refactor: 빌드 로직 개선, gradle 파일 경량화

refactor: 빌드 최적화 및 ci 수정
chore: 문서 추가
2026-09-18 15:22:30 +09:00 · 2026-09-17 15:24:31 +09:00 · 2026-09-17 15:03:36 +09:00 · 2026-09-16 20:33:19 +09:00 · 2026-09-16 18:47:15 +09:00 · 2026-09-16 18:46:34 +09:00
6616 changed files with 616507 additions and 87116 deletions
@@ -0,0 +1,41 @@
+name: Set up Java and Gradle
+description: >-
+  Installs the repository's pinned Temurin JDK, then configures Gradle through the official
+  setup-gradle action — which validates every checked-in wrapper jar and manages the Gradle cache.
+  Every Gradle job used to carry the JDK block verbatim, so the JDK patch level lived in fifty-nine
+  places; every job also carried a separate three-line wrapper-validation step, so the pinned action
+  SHA lived in forty.
+
+# Wrapper validation is INSIDE this action now.
+#
+# It could not be before, and the reason was not a GitHub limitation: .github/scripts/
+# verify-gradle-wrapper.sh read every workflow job and required it to contain, literally and in this
+# order, an `actions/checkout@` step, the exact three-field pinned wrapper-validation step, and then
+# the Gradle invocation. That literalness was the whole guard — "this job validated the wrapper" had
+# to be answerable from the workflow file alone — and it is what made the step uninlineable.
+#
+# That script is gone (it also byte-hashed all twelve workflow files, so a comment change needed a
+# hash update, while an attacker with write access would simply have updated both). The guarantee it
+# was protecting is now the official action's own: `gradle/actions/setup-gradle` validates all
+# wrapper jars by default (`validate-wrappers`, default true), and the action is pinned to a full
+# commit SHA here — which GitHub's own hardening guide calls the only immutable action reference.
+#
+# `actions/checkout` still cannot move here: a `./.github/actions/...` reference is resolved from the
+# checked-out working copy, so this file does not exist until checkout has already run. A composite
+# action cannot contain the step that makes itself readable.
+#
+# So a Gradle job is two lines — checkout, then this action.
+
+runs:
+  using: composite
+  steps:
+    - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
+      with:
+        distribution: temurin
+        java-version: "21.0.11+10"
+    # Gradle's own caching, not setup-java's `cache: gradle`. The two cache the same directory with
+    # different keys, and running both is how a job restores one cache and saves the other.
+    - uses: gradle/actions/setup-gradle@3f131e8634966bd73d06cc69884922b02e6faf92 # gradle/actions@v6
+      with:
+        build-scan-publish: false
+        cache-read-only: ${{ github.ref != 'refs/heads/main' }}
@@ -1,194 +0,0 @@
-# Current repository CI controls. This file lists only mechanisms and jobs that exist in this
-# checkout. Build/release supply-chain, image, signing, provenance, SBOM, and tag-release jobs are
-# intentionally absent until their later bounded reconstruction.
-#
-# Fields:
-#   release_blocking: true, false, or conditional
-#   mechanism: gradle-custom-task, gradle-plugin-task, contract-test, workflow-job,
-#              or delegated-pending
-#   ref: task, plugin@task, repository-relative test path below src/, or workflow job id
-#   workflow/job: canonical workflow and job that execute or represent the control
-#   execution: check (through Gradle check), explicit (named in the job), or job
-gates:
-  - id: format-lint
-    release_blocking: true
-    mechanism: gradle-plugin-task
-    ref: com.diffplug.spotless@spotlessCheck
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: unit-and-contract-tests
-    release_blocking: true
-    mechanism: gradle-plugin-task
-    ref: java@test
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: clean-architecture-dependencies
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyCleanArchitectureDependencies
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: environment-contract
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyEnvKeys
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: one-type-per-file
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyOneTypePerFile
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: readme-command-drift
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyReadmeCommands
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: trivy-suppression-governance
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyTrivyignore
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: quarantine-sunset
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyQuarantineSunset
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: public-path-snapshot
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyPublicPathSnapshot
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: explicit
-  - id: dependency-locks
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: verifyDependencyLocks
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: explicit
-  - id: architecture-contract-test
-    release_blocking: true
-    mechanism: contract-test
-    ref: app-bootstrap/src/test/java/dev/caskeleton/bootstrap/architecture/CleanArchitectureTest.java
-    workflow: ci-quality-gates.yml
-    job: quality-gates
-    execution: check
-  - id: sample-off
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: sampleOffTest
-    workflow: ci-quality-gates.yml
-    job: sample-off
-    execution: explicit
-  - id: gate-matrix-lint
-    release_blocking: true
-    mechanism: workflow-job
-    ref: gate-matrix-lint
-    workflow: ci-quality-gates.yml
-    job: gate-matrix-lint
-    execution: job
-  - id: redis-standalone
-    release_blocking: true
-    mechanism: workflow-job
-    ref: redis-standalone
-    workflow: ci-quality-gates.yml
-    job: redis-standalone
-    execution: job
-  - id: jpa-candidate-evidence
-    release_blocking: true
-    mechanism: workflow-job
-    ref: jpa-candidate-evidence
-    workflow: ci-quality-gates.yml
-    job: jpa-candidate-evidence
-    execution: job
-  - id: jpa-r2-evidence
-    release_blocking: conditional
-    mechanism: workflow-job
-    ref: jpa-r2-evidence
-    workflow: jpa-r2-evidence.yml
-    job: jpa-r2-evidence
-    execution: job
-  - id: quality-release-gate
-    release_blocking: true
-    mechanism: workflow-job
-    ref: release-gate
-    workflow: ci-quality-gates.yml
-    job: release-gate
-    execution: job
-  - id: flaky-quarantine
-    release_blocking: false
-    mechanism: workflow-job
-    ref: quarantine
-    workflow: ci-quality-gates.yml
-    job: quarantine
-    execution: job
-  - id: dependency-review
-    release_blocking: conditional
-    mechanism: workflow-job
-    ref: dependency-review
-    workflow: dependency-vulnerability.yml
-    job: dependency-review
-    execution: job
-  - id: dependency-submission
-    release_blocking: false
-    mechanism: workflow-job
-    ref: dependency-submission
-    workflow: dependency-vulnerability.yml
-    job: dependency-submission
-    execution: job
-  - id: filesystem-vulnerability-scan
-    release_blocking: true
-    mechanism: workflow-job
-    ref: trivy-fs
-    workflow: dependency-vulnerability.yml
-    job: trivy-fs
-    execution: job
-  - id: documentation-links
-    release_blocking: conditional
-    mechanism: workflow-job
-    ref: lychee
-    workflow: link-check.yml
-    job: lychee
-    execution: job
-  - id: object-storage-minio-managed-contract
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: objectStorageMinioContractTest
-    workflow: object-storage-qualification.yml
-    job: minio-managed-contract
-    execution: explicit
-  - id: poster-image-v7-migration
-    release_blocking: true
-    mechanism: gradle-custom-task
-    ref: posterImageMigrationTest
-    workflow: object-storage-qualification.yml
-    job: poster-image-v7-migration
-    execution: explicit
-  - id: object-storage-minio-managed-fault
-    release_blocking: conditional
-    mechanism: gradle-custom-task
-    ref: objectStorageMinioFaultTest
-    workflow: object-storage-qualification.yml
-    job: minio-managed-fault
-    execution: explicit
-  - id: object-storage-aws-protected-qualification
-    release_blocking: conditional
-    mechanism: delegated-pending
-    ref: approval-gate-b
-    workflow: object-storage-qualification.yml
-    job: aws-managed-common-subset
-    execution: job
@@ -3,7 +3,7 @@
 This policy is enforced by
 [`dependency-vulnerability.yml`](workflows/dependency-vulnerability.yml),
 [`dependency-review-config.yml`](dependency-review-config.yml),
-[`../.trivyignore.yaml`](../.trivyignore.yaml), `verifyTrivyignore`, CODEOWNERS, and
+[`../.trivyignore.yaml`](../.trivyignore.yaml), CODEOWNERS, and
 [`../renovate.json`](../renovate.json).

 ## Execution and platform boundary
@@ -73,7 +73,7 @@ dependencies; stale mirrors can delay detection.

 The only suppression source is repository-root `.trivyignore.yaml`. Every Trivy scan passes it
 explicitly with `--ignorefile .trivyignore.yaml`. Each future entry must contain an identifier, a
-non-empty rationale, and a future expiry no more than 90 days away. `verifyTrivyignore` validates
+non-empty rationale, and a future expiry no more than 90 days away. A CODEOWNERS reviewer validates
 the shape and expiry; CODEOWNERS plus branch protection controls who may approve the change.
 Neither control substitutes for the other.

@@ -4,8 +4,8 @@

 ## Verification

- [ ] I ran the focused test for each changed leaf.
- [ ] I ran `cd src && ./gradlew check verifyPublicPathSnapshot verifyDependencyLocks`.
+- [ ] I ran focused `:<changed-leaf>:check` tasks for the modules I changed.
+- [ ] I ran `cd src && ./gradlew architectureCheck verifyPublicPathSnapshot verifyDependencyLocks` when the change touched repository structure, dependencies, or public paths.
 - [ ] I did not add an unregistered production module dependency.
 - [ ] Dependency changes include refreshed `gradle.lockfile` files and a strict-lock verification.
 - [ ] Trivy suppressions include an owner-reviewed reason and an expiry within 90 days.
@@ -1,202 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-readonly SCRIPT_DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd -P)"
-readonly REPO_ROOT="$(git -C "${SCRIPT_DIR}" rev-parse --show-toplevel)"
-readonly EXPECTED_SCRIPT_DIR="$(cd -- "${REPO_ROOT}/.github/scripts" && pwd -P)"
-readonly MATRIX="${REPO_ROOT}/.github/ci-gate-matrix.yml"
-readonly EXPECTED_GATE_COUNT=26
-
-if [[ "${SCRIPT_DIR}" != "${EXPECTED_SCRIPT_DIR}" ]]; then
-  printf '::error::gate-matrix-lint: script resolved outside the repository .github/scripts directory\n' >&2
-  exit 1
-fi
-if [[ ! -f "${MATRIX}" ]]; then
-  printf '::error::gate-matrix-lint: missing %s\n' "${MATRIX}" >&2
-  exit 1
-fi
-
-records="$(
-  awk '
-    function flush() {
-      if (id != "") {
-        printf "%s\t%s\t%s\t%s\t%s\t%s\t%s\n", id, blocking, mechanism, ref, workflow, job, execution
-      }
-    }
-    /^[[:space:]]*-[[:space:]]+id:[[:space:]]*/ {
-      flush()
-      id=$0
-      sub(/^[[:space:]]*-[[:space:]]+id:[[:space:]]*/, "", id)
-      blocking=mechanism=ref=workflow=job=execution=""
-      next
-    }
-    /^[[:space:]]+release_blocking:[[:space:]]*/ {
-      blocking=$0
-      sub(/^[[:space:]]+release_blocking:[[:space:]]*/, "", blocking)
-      next
-    }
-    /^[[:space:]]+mechanism:[[:space:]]*/ {
-      mechanism=$0
-      sub(/^[[:space:]]+mechanism:[[:space:]]*/, "", mechanism)
-      next
-    }
-    /^[[:space:]]+ref:[[:space:]]*/ {
-      ref=$0
-      sub(/^[[:space:]]+ref:[[:space:]]*/, "", ref)
-      next
-    }
-    /^[[:space:]]+workflow:[[:space:]]*/ {
-      workflow=$0
-      sub(/^[[:space:]]+workflow:[[:space:]]*/, "", workflow)
-      next
-    }
-    /^[[:space:]]+job:[[:space:]]*/ {
-      job=$0
-      sub(/^[[:space:]]+job:[[:space:]]*/, "", job)
-      next
-    }
-    /^[[:space:]]+execution:[[:space:]]*/ {
-      execution=$0
-      sub(/^[[:space:]]+execution:[[:space:]]*/, "", execution)
-      next
-    }
-    END { flush() }
-  ' "${MATRIX}"
-)"
-
-declare -A seen_ids=()
-declare -a failures=()
-total=0
-verified=0
-delegated=0
-
-job_body() {
-  local workflow_file="$1"
-  local job_id="$2"
-  awk -v target="${job_id}" '
-    $0 ~ "^  " target ":[[:space:]]*$" { inside=1; print; next }
-    inside && $0 ~ "^  [A-Za-z0-9_-]+:[[:space:]]*$" { exit }
-    inside { print }
-  ' "${workflow_file}"
-}
-
-while IFS=$'\t' read -r id blocking mechanism ref workflow job execution; do
-  [[ -z "${id}" ]] && continue
-  total=$((total + 1))
-
-  if [[ -n "${seen_ids[${id}]:-}" ]]; then
-    failures+=("duplicate gate id '${id}'")
-  fi
-  seen_ids["${id}"]=1
-
-  if [[ -z "${blocking}" || -z "${mechanism}" || -z "${ref}" || -z "${workflow}" \
-        || -z "${job}" || -z "${execution}" ]]; then
-    failures+=("gate '${id}' has an empty required field")
-    continue
-  fi
-  if [[ ! "${blocking}" =~ ^(true|false|conditional)$ ]]; then
-    failures+=("gate '${id}' has invalid release_blocking '${blocking}'")
-  fi
-  if [[ ! "${workflow}" =~ ^[A-Za-z0-9._-]+\.ya?ml$ || ! "${job}" =~ ^[A-Za-z0-9_-]+$ ]]; then
-    failures+=("gate '${id}' has an unsafe workflow or job identifier")
-    continue
-  fi
-
-  workflow_file="${REPO_ROOT}/.github/workflows/${workflow}"
-  if [[ ! -f "${workflow_file}" ]]; then
-    failures+=("gate '${id}' references missing workflow '.github/workflows/${workflow}'")
-    continue
-  fi
-  if ! grep -Eqs -- "^[[:space:]]{2}${job}:[[:space:]]*$" "${workflow_file}"; then
-    failures+=("gate '${id}' references missing job '${job}' in '${workflow}'")
-    continue
-  fi
-
-  case "${mechanism}" in
-    gradle-custom-task)
-      if ! grep -RqsE -- "tasks\\.register\\(['\"]${ref}['\"]" "${REPO_ROOT}/src" \
-          --include='build.gradle'; then
-        failures+=("gate '${id}' references unregistered Gradle task '${ref}'")
-        continue
-      fi
-      ;;
-    gradle-plugin-task)
-      plugin="${ref%@*}"
-      task="${ref#*@}"
-      if [[ "${plugin}" == "${ref}" || -z "${task}" ]]; then
-        failures+=("gate '${id}' must use plugin@task for gradle-plugin-task")
-        continue
-      fi
-      if ! grep -RqsE -- "(id|apply plugin:)[[:space:]]+['\"]${plugin}['\"]" "${REPO_ROOT}/src" \
-          --include='build.gradle'; then
-        failures+=("gate '${id}' references unapplied Gradle plugin '${plugin}'")
-        continue
-      fi
-      ;;
-    contract-test)
-      if [[ "${ref}" == /* || "${ref}" == *".."* || ! -f "${REPO_ROOT}/src/${ref}" ]]; then
-        failures+=("gate '${id}' references missing or unsafe contract test 'src/${ref}'")
-        continue
-      fi
-      ;;
-    workflow-job)
-      if [[ "${ref}" != "${job}" ]]; then
-        failures+=("gate '${id}' workflow-job ref '${ref}' must equal job '${job}'")
-        continue
-      fi
-      ;;
-    delegated-pending)
-      delegated=$((delegated + 1))
-      printf "gate '%s': explicitly delegated-pending\n" "${id}"
-      continue
-      ;;
-    *)
-      failures+=("gate '${id}' has unknown mechanism '${mechanism}'")
-      continue
-      ;;
-  esac
-
-  case "${execution}" in
-    check)
-      if ! job_body "${workflow_file}" "${job}" | grep -Eqs -- '\./gradlew[[:space:]]+check([[:space:]]|$)'; then
-        failures+=("gate '${id}' expects Gradle check in job '${job}'")
-        continue
-      fi
-      if [[ "${mechanism}" == "gradle-custom-task" ]] \
-          && ! grep -RqsE -- "dependsOn.*named\\(['\"]${ref}['\"]\\)" "${REPO_ROOT}/src" \
-            --include='build.gradle'; then
-        failures+=("gate '${id}' task '${ref}' exists but is not wired into Gradle check")
-        continue
-      fi
-      ;;
-    explicit)
-      if ! job_body "${workflow_file}" "${job}" | grep -Fqs -- "${ref}"; then
-        failures+=("gate '${id}' task '${ref}' is not explicit in job '${job}'")
-        continue
-      fi
-      ;;
-    job)
-      ;;
-    *)
-      failures+=("gate '${id}' has unknown execution '${execution}'")
-      continue
-      ;;
-  esac
-
-  verified=$((verified + 1))
-done <<< "${records}"
-
-if (( total != EXPECTED_GATE_COUNT )); then
-  failures+=("matrix has ${total} gates; expected ${EXPECTED_GATE_COUNT}")
-fi
-
-printf 'gate-matrix-lint: %d gates, %d verified, %d delegated-pending\n' \
-  "${total}" "${verified}" "${delegated}"
-if (( ${#failures[@]} > 0 )); then
-  printf '::error::gate-matrix-lint: %d drift(s) found\n' "${#failures[@]}" >&2
-  for failure in "${failures[@]}"; do
-    printf '  - %s\n' "${failure}" >&2
-  done
-  exit 1
-fi
-printf 'gate-matrix-lint: OK\n'
@@ -0,0 +1,63 @@
+name: reusable-gradle
+
+# One place that knows how a Gradle job starts.
+#
+# Every job in this repository opened with the same preamble: checkout, a three-line pinned
+# wrapper-validation step, then the JDK/cache action. The wrapper step is gone (setup-gradle
+# validates wrappers itself), and this workflow removes the rest of the repetition for the jobs whose
+# only variation is the Gradle command they run.
+#
+# Jobs that need service containers, a matrix, artifact uploads or per-job env stay written out with
+# `./.github/actions/setup-gradle-java`, because expressing those through `workflow_call` inputs
+# means encoding YAML inside strings — which is how a "shared" workflow becomes less readable than
+# the duplication it replaced.
+
+on:
+  workflow_call:
+    inputs:
+      tasks:
+        description: The Gradle task list, whitespace-separated.
+        required: true
+        type: string
+      gradle-args:
+        description: Flags appended after the task list.
+        required: false
+        type: string
+        default: "--stacktrace"
+      working-directory:
+        description: Directory the wrapper is invoked from.
+        required: false
+        type: string
+        default: src
+      timeout-minutes:
+        required: false
+        type: number
+        default: 30
+      continue-on-error:
+        description: Run the job as an advisory signal rather than a gate.
+        required: false
+        type: boolean
+        default: false
+
+permissions:
+  contents: read
+
+jobs:
+  gradle:
+    runs-on: ubuntu-latest
+    timeout-minutes: ${{ inputs.timeout-minutes }}
+    continue-on-error: ${{ inputs.continue-on-error }}
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run ${{ inputs.tasks }}
+        working-directory: ${{ inputs.working-directory }}
+        env:
+          GRADLE_TASKS: ${{ inputs.tasks }}
+          GRADLE_ARGS: ${{ inputs.gradle-args }}
+        run: |
+          set -euo pipefail
+          # Word-split on purpose: both inputs are task/flag lists. They come from this repository's
+          # own workflow files, never from a pull request.
+          # shellcheck disable=SC2086
+          ./gradlew ${GRADLE_TASKS} ${GRADLE_ARGS}
@@ -1,5 +1,14 @@
 name: ci-quality-gates

+# The pull-request gate. Everything here blocks a merge.
+#
+# The job list used to include `gate-matrix-lint`, which ran .github/scripts/verify-gate-matrix.sh
+# against .github/ci-gate-matrix.yml: a 1,025-line register of all 107 CI controls, checked for
+# consistency against the Gradle task graph and this workflow by a 568-line shell script, which was
+# itself checked by contract tests in :app-bootstrap. Adding one check meant editing Gradle, a
+# workflow, the matrix, the verifier's expectations and a Java test. The information was already in
+# the task graph and the job graph; the matrix was a third copy that had to be kept equal to both.
+
 on:
  pull_request:
  push:
@@ -33,90 +42,97 @@ jobs:
            echo "::error::${snapshot} exists locally but is not committed."
            exit 1
          fi
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
-      - name: Check quality, public paths, and dependency locks
+      - uses: ./.github/actions/setup-gradle-java
+      # `ci`, not `check`. A leaf's `check` is that leaf's — compile, its tests, Spotless, Checkstyle
+      # and Error Prone — and the repository-wide gates are named tasks of their own:
+      #   ci = every leaf check + architectureCheck + qualityCheck + configContractCheck + qualificationCheck
+      # so CI runs strictly more than it used to while `./gradlew :domain-core:check` runs strictly
+      # less.
+      - name: Run the pull-request gate
        working-directory: src
-        run: ./gradlew check verifyPublicPathSnapshot verifyDependencyLocks --no-daemon --stacktrace
-
-  sample-off:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
-      - name: Verify the application without the sample fixture
+        run: ./gradlew :ci :verifyPublicPathSnapshot :verifyDependencyLocks --warning-mode=fail --stacktrace
+      # Named as its own step because nothing else runs it: `check` does not depend on
+      # graphqlStableTest, so the lane's required-class guard — the check that its module-boundary
+      # suite has not silently stopped being discovered — would protect nothing in CI.
+      - name: Qualify the GraphQL Stable lane
        working-directory: src
-        run: ./gradlew :app-bootstrap:sampleOffTest verifyCleanArchitectureDependencies --no-daemon --stacktrace
+        run: ./gradlew :adapter:inbound:graphql:graphqlStableTest --stacktrace
+      - name: Qualify opt-in inbound transports without skips
+        working-directory: src
+        run: ./gradlew :conditionalTransportQualification --stacktrace

-  gate-matrix-lint:
+  build-logic:
+    # Included-build tests are independent of the main project task graph. Running them as a
+    # separate blocking job keeps plugin TestKit work off the quality-gates critical path.
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - name: Verify the gate matrix against the repository
-        run: bash .github/scripts/verify-gate-matrix.sh
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Test the build-logic convention plugins
+        working-directory: src
+        run: ./gradlew -p build-logic test --stacktrace

-  redis-standalone:
+  redis-sdk:
+    # Milestone A of the Redis wrapper/typed API plan: policy catalog, typed API parity, permit
+    # provenance, connection isolation, and the executor guard. There is no real-server lane yet.
+    #
+    # `verifyConfigurationPropertiesProcessor` used to be in this list. It is deleted: the parity it
+    # enforced — a leaf declares Spring's configuration processor exactly when it owns
+    # @ConfigurationProperties — is now what applying `ca.spring-config` means.
+    # `verifyEnvKeys` is no longer named here either; it belongs to :app-bootstrap and runs through
+    # `configContractCheck`, which the quality-gates job covers.
+    uses: ./.github/workflows/_reusable-gradle.yml
+    with:
+      tasks: ":shared-contract:edgeRateLimitContractTest"
+
+  optional-platforms:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
-      - name: Verify standalone Redis policy, provider, and composition contracts
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Verify the optional gRPC platform build
+        working-directory: src
+        run: ./gradlew -p optional-platforms ci --stacktrace
+
+  configuration-cache:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Store configuration cache for the everyday core build
        working-directory: src
        run: >-
          ./gradlew
-          :application-core:redisPolicyContractTest
-          :shared-contract:edgeRateLimitContractTest
-          :adapter:outbound:cache-redis:check
-          :app-bootstrap:redisCompositionTest
-          verifyCleanArchitectureDependencies
-          verifyEnvKeys
-          verifyPublicPathSnapshot
-          verifyConfigurationPropertiesProcessor
-          --no-daemon --stacktrace
+          :domain-core:check
+          :application-core:check
+          --configuration-cache
+          --configuration-cache-problems=fail
+          --stacktrace
+      - name: Require configuration-cache reuse
+        working-directory: src
+        shell: bash
+        run: |
+          set -euo pipefail
+          output="$({ ./gradlew :domain-core:check :application-core:check \
+            --configuration-cache --configuration-cache-problems=fail --stacktrace; } 2>&1)"
+          printf '%s\n' "${output}"
+          grep -Fq 'Reusing configuration cache.' <<<"${output}" || {
+            echo '::error::Gradle did not reuse the configuration cache on the second identical build.'
+            exit 1
+          }

  jpa-candidate-evidence:
    runs-on: ubuntu-latest
    timeout-minutes: 20
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
+      - uses: ./.github/actions/setup-gradle-java
      - name: Produce zero-skip JPA candidate manifests
        working-directory: src
        run: >-
          ./gradlew
          :adapter:outbound:persistence-jpa:verifyJpaCandidateEvidence
-          --no-daemon
+
          --stacktrace
      - name: Retain content-addressed JPA candidate manifests
        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # actions/upload-artifact@v7.0.1
@@ -126,50 +142,46 @@ jobs:
          if-no-files-found: error
          retention-days: 14

-  # Advisory only. Quarantine expiry/drift remains blocking through verifyQuarantineSunset in check.
+  # Advisory. The quarantine bucket runs so a flaky test is still executed and reported; it never
+  # blocks. The 14-day sunset registry that used to make an expired quarantine entry a build failure
+  # is gone — it was a 250-line YAML-and-Java parser guarding a registry with zero entries.
  quarantine:
-    runs-on: ubuntu-latest
-    continue-on-error: true
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
-      - name: Run quarantined tests as an advisory signal
-        working-directory: src
-        run: ./gradlew quarantineTest --no-daemon
+    uses: ./.github/workflows/_reusable-gradle.yml
+    with:
+      tasks: ":quarantineTest"
+      gradle-args: "--stacktrace"
+      continue-on-error: true

  release-gate:
    needs:
      - quality-gates
-      - sample-off
-      - gate-matrix-lint
-      - redis-standalone
+      - build-logic
+      - redis-sdk
      - jpa-candidate-evidence
+      - optional-platforms
+      - configuration-cache
    if: always()
    runs-on: ubuntu-latest
+    permissions:
+      contents: read
    steps:
      - name: Require every current blocking job to succeed
        env:
          QUALITY_RESULT: ${{ needs.quality-gates.result }}
-          SAMPLE_OFF_RESULT: ${{ needs.sample-off.result }}
-          MATRIX_RESULT: ${{ needs.gate-matrix-lint.result }}
-          REDIS_RESULT: ${{ needs.redis-standalone.result }}
+          BUILD_LOGIC_RESULT: ${{ needs.build-logic.result }}
+          REDIS_RESULT: ${{ needs.redis-sdk.result }}
          JPA_CANDIDATE_RESULT: ${{ needs.jpa-candidate-evidence.result }}
+          OPTIONAL_PLATFORMS_RESULT: ${{ needs.optional-platforms.result }}
+          CONFIGURATION_CACHE_RESULT: ${{ needs.configuration-cache.result }}
        run: |
          set -euo pipefail
          for result in \
            "${QUALITY_RESULT}" \
-            "${SAMPLE_OFF_RESULT}" \
-            "${MATRIX_RESULT}" \
+            "${BUILD_LOGIC_RESULT}" \
            "${REDIS_RESULT}" \
-            "${JPA_CANDIDATE_RESULT}"; do
+            "${JPA_CANDIDATE_RESULT}" \
+            "${OPTIONAL_PLATFORMS_RESULT}" \
+            "${CONFIGURATION_CACHE_RESULT}"; do
            if [[ "${result}" != "success" ]]; then
              echo "::error::release-gate: required job result was ${result}"
              exit 1
@@ -35,15 +35,7 @@ jobs:
      contents: write
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
+      - uses: ./.github/actions/setup-gradle-java
      - name: Submit the resolved Gradle dependency graph
        uses: gradle/actions/dependency-submission@748248ddd2a24f49513d8f472f81c3a07d4d50e1 # gradle/actions@v4.4.4
        with:
@@ -176,7 +168,10 @@ jobs:
            trivy-kev.json | sort -u > found-cves.txt
          jq -r '.vulnerabilities[]?.cveID | select(type == "string")' \
            kev.json | sort -u > kev-cves.txt
-          hits="$(comm -12 found-cves.txt kev-cves.txt || true)"
+          # No `|| true`. comm exits non-zero only when it cannot read or order its inputs, and
+          # swallowing that would have turned an unreadable CVE list into an empty intersection and
+          # printed "no catalog match" — a KEV cross-check that passes because it never ran.
+          hits="$(comm -12 found-cves.txt kev-cves.txt)"
          if [[ -n "${hits}" ]]; then
            echo "::error::CISA KEV-listed vulnerability found regardless of CVSS:"
            printf '%s\n' "${hits}"
@@ -0,0 +1,122 @@
+name: fileserver-certification
+
+# The certification a release must clear. Its job list is deliberately the same shape as the support
+# matrix: nothing may be advertised at a support level whose evidence job is absent here.
+#
+# Named "certification", not "release", and the name is the point. This workflow proves a storage
+# topology, a support matrix and a telemetry redaction claim. It deploys nothing and holds no cluster
+# credential. Calling it `fileserver-release.yml` read as if GitHub Actions released the fileserver,
+# which is the CI/CD boundary this repository has now fixed in docs/ci-cd/boundary.md: GitHub Actions
+# tests, scans and publishes artifacts; Argo CD deploys.
+#
+# It used to be workflow_dispatch only, which made that sentence false: the four jobs below are the
+# only place the fileserver support matrix, the PVC manifest and the telemetry redaction proof are
+# checked, and a release tag reached none of them unless somebody remembered to press a button.
+#
+# `v*` is the only release tag. The adapter-scoped `fileserver-v*` pattern is gone: this repository
+# has one deployable unit (app-bootstrap), so an adapter-scoped tag could only ever run a subset of
+# the release gates and call the result a release — the tag-namespace split that release.yml exists
+# to end.
+#
+# These four jobs stay in their own file, and not in release.yml, for one mechanical reason:
+# FileserverDocumentationCoverageTest reads job ids out of `.github/workflows/fileserver-*.yml` and
+# requires every `fileserver-...` job docs/fileserver/support-matrix.md names to be defined in one
+# of them. Renaming the file or moving these jobs needs that document changed in the same change.
+
+on:
+  push:
+    tags:
+      - "v*"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: false
+
+jobs:
+  fileserver-full-verification:
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the architecture-wide dependency and module verification
+        working-directory: src
+        run: >-
+          ./gradlew
+          :verifyCleanArchitectureDependencies
+
+          --stacktrace
+      - name: Run the complete fileserver suite across every leaf
+        working-directory: src
+        run: >-
+          ./gradlew
+          :application-core:check
+          :adapter:inbound:web:check
+          :adapter:outbound:fileserver:check
+
+          --stacktrace
+
+  fileserver-documentation-gate:
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Prove every support claim maps to a job and every endpoint is documented
+        working-directory: src
+        run: >-
+          ./gradlew
+          :app-bootstrap:test --tests '*FileserverDocumentationCoverageTest'
+
+          --stacktrace
+
+  fileserver-pvc-certification:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      # This job checks the manifest, and only the manifest. It deliberately does not apply anything
+      # to a cluster.
+      #
+      # There used to be a second step here that applied the job to a release cluster when
+      # secrets.FILESERVER_PVC_KUBECONFIG was set and `exit 0`-ed with a ::warning:: when it was
+      # not. With no secret configured — which is every fork of this template and was this
+      # repository — the step printed a warning and the job went green under the name
+      # "fileserver-pvc-certification", so a release read as ReadWriteOnce-certified against a
+      # cluster nothing had ever touched. It also wrote a `certified` output that no job, step or
+      # script in this repository read.
+      #
+      # The cluster result comes from an operator running infra/fileserver/kubernetes/
+      # pvc-certification-job.yaml against a real cluster and recording it in
+      # docs/fileserver/storage-certification.md. That is registered as
+      # docs/fileserver/storage-certification.md, and the absence of a cluster result is stated
+      # there rather than hidden behind a green check.
+      - name: Check the certification manifest still says what the claim depends on
+        run: |
+          set -euo pipefail
+          manifest=infra/fileserver/kubernetes/pvc-certification-job.yaml
+          test -f "$manifest"
+          grep -q 'kind: PersistentVolumeClaim' "$manifest"
+          grep -q 'kind: Job' "$manifest"
+          # ReadWriteMany is explicitly not claimed; a manifest that quietly widened the access
+          # mode would certify a topology the support matrix says is uncertified.
+          grep -q 'ReadWriteOnce' "$manifest"
+          ! grep -q 'ReadWriteMany' "$manifest"
+
+  fileserver-sensitive-telemetry-scan:
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Prove telemetry carries no filename, path, or raw identifier
+        working-directory: src
+        run: >-
+          ./gradlew
+          :application-core:test --tests '*FileserverObservabilityTest'
+
+          --stacktrace
@@ -0,0 +1,88 @@
+name: fileserver-nightly
+
+# The environments that cannot run on every pull request: a real network filesystem, a foreign
+# filesystem, and the long-running fault matrices. They are nightly rather than skipped because a
+# green pull-request run is not certification of any of them.
+
+on:
+  workflow_dispatch:
+  schedule:
+    - cron: '0 18 * * *'
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}
+  cancel-in-progress: false
+
+jobs:
+  fileserver-nfs-ambiguity:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    env:
+      FILESERVER_NFS_TESTS: "true"
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Start the NFSv4 certification environment
+        run: docker compose -f infra/fileserver/nfs/compose.yml up -d --wait
+      - name: Run the network-filesystem ambiguity suite
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:fileserver:test --tests '*NfsAmbiguityIntegrationTest'
+
+          --stacktrace
+      - name: Tear down the NFS environment
+        if: always()
+        run: docker compose -f infra/fileserver/nfs/compose.yml down -v
+
+  fileserver-process-kill-matrix:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the crash matrix and reconciliation suites
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:fileserver:test --tests '*CrashRecoveryMatrixTest'
+          :application-core:test --tests '*FileReconciliationServiceTest'
+          --rerun-tasks
+
+          --stacktrace
+
+  fileserver-large-file-performance:
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the large-file and slow-client suites under a constrained heap
+        working-directory: src
+        env:
+          GRADLE_OPTS: -Xmx512m
+        run: >-
+          ./gradlew
+          :adapter:outbound:fileserver:test --tests '*LargeFileBoundedMemoryTest'
+          :adapter:outbound:fileserver:test --tests '*LocalAppendMemoryTest'
+          --rerun-tasks
+
+          --stacktrace
+
+  fileserver-multi-instance-lease:
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Prove no run commits bytes from a stale lease
+        working-directory: src
+        run: >-
+          ./gradlew
+          :application-core:test --tests '*MultiInstanceWriterLeaseTest'
+          --rerun-tasks
+
+          --stacktrace
@@ -0,0 +1,113 @@
+name: fileserver-pr
+
+# Every claim in docs/fileserver/support-matrix.md that says "Stable" is backed by a job here.
+# A support level with no job behind it is a marketing claim, not an engineering one, and
+# DocumentationCoverageTest fails the build when the two drift apart.
+
+on:
+  workflow_dispatch:
+  pull_request:
+    paths:
+      - 'src/application-core/src/**/fileserver/**'
+      - 'src/adapter/inbound/web/src/**/fileserver/**'
+      - 'src/adapter/outbound/fileserver/**'
+      - 'src/adapter/outbound/persistence-jpa/src/**/fileserver/**'
+      - 'src/app-bootstrap/src/**/fileserver/**'
+      - 'docs/fileserver/**'
+      # The capability is not only its Java files. A change to the bound settings, the shipped
+      # environment, the registry that documents it, or the container that has to give it a
+      # writable volume changes how it behaves at runtime just as surely — and those were the
+      # exact files that could previously ship unverified.
+      - 'src/app-bootstrap/src/main/resources/application.yml'
+      - 'src/.env'
+      - 'docs/registries/env-keys.yaml'
+      - 'src/Dockerfile'
+      - 'docker-compose.yml'
+      - 'infra/fileserver/nginx/**'
+      - 'infra/fileserver/kubernetes/**'
+      - 'infra/fileserver/nfs/**'
+      - '.github/workflows/fileserver-pr.yml'
+      # Every Gradle job here installs its toolchain through this composite action, so a change to
+      # it changes what this gate runs.
+      - '.github/actions/setup-gradle-java/action.yml'
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  fileserver-unit-and-architecture:
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the fileserver application and architecture suites
+        working-directory: src
+        run: >-
+          ./gradlew
+          :application-core:test
+          :app-bootstrap:test --tests 'dev.caskeleton.bootstrap.architecture.*' --tests '*Fileserver*'
+
+          --stacktrace
+
+  fileserver-local-ext4-contract:
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Certify the local content store against the shared contract
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:fileserver:test
+
+          --stacktrace
+
+  fileserver-http-contract:
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the servlet and reactive transport contracts
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:web:test
+
+          --stacktrace
+
+  fileserver-security-suite:
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the path, filename, range, and problem-detail hardening suite
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:web:test --tests '*FileserverHardeningContractTest'
+          :adapter:outbound:fileserver:test --tests '*PhysicalPathResolverTest'
+
+          --stacktrace
+
+  fileserver-bounded-memory:
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Prove transfer cost does not scale with file size
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:fileserver:test --tests '*LargeFileBoundedMemoryTest'
+          :adapter:inbound:web:test --tests '*DataBufferReleaseTest'
+
+          --stacktrace
@@ -0,0 +1,232 @@
+name: integration-main
+
+# Stage 2: is the merged state healthy.
+#
+# The question this stage answers is different from stage 1's. Stage 1 asks whether a diff is safe
+# and blocks a merge; stage 2 asks whether main is healthy and does not — the merge has already
+# happened. That difference is the point, and it is what lets a control exist without being an
+# obstacle: a gate here still fails loudly, it just fails after the thing it is reporting on.
+#
+# Two kinds of work live here.
+#
+# 1. The lanes that need a machine that is not simultaneously compiling something else — load,
+#    abuse, graceful shutdown, TCP fault injection, resource bounds. They were web-nightly.yml and
+#    httpclient-nightly.yml, two module-shaped files whose only real difference was the cadence they
+#    shared. They now run on every push to main as well as nightly, which is strictly more often
+#    than before.
+#
+# 2. Lanes that were registered in Gradle and invoked by nothing. Ten Gradle tasks — six MongoDB
+#    container lanes, app-bootstrap's Testcontainers `integrationTest`, and the three messaging
+#    evidence tasks that `verifyMessagingContracts` reaches — existed, failed closed, and executed
+#    in no workflow. A lane nobody runs is not coverage; it is a file that looks like coverage. They
+#    are here rather than in stage 1 because every one of them either starts containers or re-runs
+#    suites the PR gate already covers, and the pull-request budget is minutes for the whole gate.
+#
+# What is deliberately NOT here: the web and WebSocket "Advanced capability" nightly lanes that used
+# to exist as web-advanced-nightly.yml and websocket-advanced-nightly.yml. Both leaves' build files
+# say it outright — "They also run inside `test`, deliberately ... excluding them from the PR gate to
+# make this lane look meaningful would mean the PR gate stopped covering a fifth of the leaf" — so
+# `webAdvancedTest` and `websocketAdvancedTest` select tagged tests that `:<leaf>:test` already runs,
+# and `:<leaf>:test` runs inside the root `check` on every pull request and every push to main. The
+# strict lanes themselves survive in release.yml, where their fail-on-nothing-discovered guard is
+# worth a job.
+
+on:
+  push:
+    branches: ["main"]
+  schedule:
+    # 03:00 UTC. Late enough that the day's merges are in, early enough that a failure is triaged
+    # before the next working day starts.
+    - cron: '0 3 * * *'
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: false
+
+# The documentation-drift gates that used to run here are gone rather than demoted.
+#
+# They were four hand-written parsers: README shell blocks compared against the Gradle task graph,
+# runbook identifiers compared against every declared Java type, a leaf count written in prose
+# compared against the registry, and a Markdown table compared against the declared source sets.
+# Each was a custom parser for a file format nobody controls, and each made a documentation edit a
+# precondition for a build. A stale sentence is a defect, but it is not one a build can be failed
+# for, and link-check.yml already answers the one documentation question with a stable machine
+# answer: does this link resolve.
+
+jobs:
+  # Load, abuse and graceful shutdown measure behaviour that degrades gradually rather than breaking
+  # outright — which is exactly the kind of regression a per-PR gate never catches.
+  web-load-abuse-and-shutdown:
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the load, abuse and shutdown lanes on every container
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:web:test
+          :adapter:inbound:web:webJettyCompatTest
+          :adapter:inbound:web:webFluxContractTest
+
+          --stacktrace
+      - name: Publish the test reports
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: web-integration-reports
+          path: src/adapter/inbound/web/build/reports/tests/
+          if-no-files-found: warn
+
+  # Needs a container runtime and real time (design §29). Separated from the per-PR gate rather than
+  # made optional inside it: a lane that cannot run here fails, it does not skip.
+  httpclient-fault-injection:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Inject TCP faults against a real upstream
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:httpclient:httpClientFailureInjectionTest
+
+          --stacktrace
+
+  httpclient-performance:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    env:
+      # A project property rather than a command-line flag, so the run command stays a plain,
+      # verifiable task invocation while the machine-dependent bounds are still asserted.
+      GRADLE_OPTS: -Dorg.gradle.project.performance.assertions.enabled=true
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Certify pool, streaming, retry, and rotation bounds
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:httpclient:httpClientPerformanceTest
+
+          --stacktrace
+
+  httpclient-http3-experimental:
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    # Experimental by design (D-08): the result is reported, never used to block a merge. Registered
+    # advisory so that "this job cannot fail the
+    # build" is written down rather than inferred from a field two hundred lines into a workflow.
+    continue-on-error: true
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Exercise the experimental HTTP/3 opt-in
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:httpclient:test
+          -Phttp3.tests.enabled=true
+
+          --stacktrace
+
+  # The six Docker-backed MongoDB lanes. Until now they ran in no workflow at all: the leaf excludes
+  # every one of their tags from `test` (build.gradle "Docker-backed lanes are excluded from the
+  # default unit run"), `check` gains only the hermetic `mongoStableContractTest`, and the only thing
+  # that named them was scripts/verify-mongodb-platform.sh, which nothing in .github invokes. Six
+  # lanes that fail closed without Docker, and no machine with Docker was ever asked to run them.
+  #
+  # Stage 2 rather than stage 1 because each lane starts real MongoDB containers — mongo:8.0.16,
+  # mongo:7.0.28 and a Toxiproxy in front of a three-node replica set. That is minutes per lane, and
+  # the pull-request budget is minutes for the whole gate.
+  #
+  # One single-line `./gradlew <task>` step per lane, not one folded command running six, because
+  # These command lines name each lane explicitly so that a lane which stops being invoked is
+  # actually executed — a folded command would leave six matrix rows unverifiable. It also means a
+  # red replica-set lane does not hide the compatibility lane behind it.
+  mongo-container-lanes:
+    runs-on: ubuntu-latest
+    timeout-minutes: 90
+    env:
+      # Reuse would hand the failover lane a replica set another lane had already faulted.
+      TESTCONTAINERS_REUSE_ENABLE: "false"
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Single-node replica set contract lane
+        working-directory: src
+        run: ./gradlew :adapter:outbound:persistence-mongo:mongoReplicaSetTest --stacktrace
+      - name: Run MongoDB integration lanes
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-mongo:mongoReplicaSetTest
+          :adapter:outbound:persistence-mongo:mongoFailoverTest
+          :adapter:outbound:persistence-mongo:mongoMigrationTest
+          :adapter:outbound:persistence-mongo:mongoCompatibilityTest
+          :adapter:outbound:persistence-mongo:mongoSecurityIntegrationTest
+          :adapter:outbound:persistence-mongo:mongoPerformanceTest
+          --stacktrace
+      - name: Publish the MongoDB lane reports
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: mongo-lane-reports
+          path: src/adapter/outbound/persistence-mongo/build/reports/tests/
+          if-no-files-found: warn
+
+  # The messaging contract evidence DAG. `verifyMessagingContracts` is the root of a chain that ran
+  # nowhere: it depends on four production qualification tasks (application-core,
+  # shared-contract and two in adapter:outbound:messaging), each of which depends on
+  # `prepareMessagingContractEvidence`; it is finalizedBy
+  # `validateMessagingContractsEvidenceManifestSchema`; and it depends on
+  # `validateMessagingJsonSchemaV1EvidenceManifestSchema`, which depends on
+  # `verifyMessagingJsonSchemaV1`. Strict qualification tasks are registered outside `check` by
+  # design (ca.strict-qualification.gradle), so none of the seven was reachable from any workflow.
+  #
+  # The schema validators are the part that matters. They re-read the manifest bytes the run just
+  # wrote and validate them against config/messaging/evidence/build-evidence-manifest-v1.schema.json
+  # — a manifest that claims a qualification nobody executed is exactly the failure they exist to
+  # catch, and until now nothing executed them either.
+  #
+  # Stage 2 rather than stage 1: no containers, but it runs four qualification suites across three
+  # leaves plus two JavaExec validators, and the tests it re-runs are already inside the PR gate's
+  # `check`. What this job adds is the evidence manifest, which is a main-branch artifact.
+  messaging-contract-evidence:
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Qualify the messaging contract, catalog, binding and schema evidence
+        working-directory: src
+        run: ./gradlew :verifyMessagingContracts --stacktrace
+      - name: Publish the messaging evidence manifest
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: messaging-contract-evidence
+          path: src/build/messaging-evidence/
+          if-no-files-found: warn
+
+  # app-bootstrap's Testcontainers lane. The leaf gave it a source set of its own precisely so that
+  # `./gradlew :app-bootstrap:test` would not require a Docker daemon — and the consequence nobody
+  # closed is that a source set outside `test` is also outside `check`, so the real-PostgreSQL
+  # outbox and idempotency contracts compiled on every build and executed on none.
+  bootstrap-integration:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    env:
+      TESTCONTAINERS_REUSE_ENABLE: "false"
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the real-PostgreSQL integration contracts
+        working-directory: src
+        run: ./gradlew :app-bootstrap:integrationTest --stacktrace
@@ -0,0 +1,75 @@
+name: jpa-next
+
+# Advisory compatibility probes for future JPA/Hibernate/PostgreSQL majors. These targets are not
+# resolved by the current build, so the artifact records NOT_EXECUTABLE rather than implying that a
+# green policy test is compatibility evidence.
+
+on:
+  workflow_dispatch:
+  schedule:
+    - cron: '0 5 * * 1'
+
+permissions:
+  contents: read
+
+jobs:
+  compatibility:
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - id: jakarta-persistence-4
+            target: Jakarta Persistence 4
+            coordinate: jakarta.persistence:jakarta.persistence-api:4.x
+            test: "*CompatibilityLaneDefinitionTest"
+            reason: the JPA 4 API is not on any configuration this build resolves, so nothing has been compiled against it
+          - id: hibernate-8
+            target: Hibernate 8
+            coordinate: org.hibernate.orm:hibernate-core:8.x
+            test: "*HibernateCompatibilityPolicyTest"
+            reason: Hibernate 8 is not resolvable from this build, so nothing has been compiled or run against it
+          - id: postgresql-19
+            target: PostgreSQL 19
+            coordinate: postgres:19-alpine
+            test: "*ExperimentalPromotionGateTest"
+            reason: no PostgreSQL 19 image is published yet, so no container of that major has ever been started by this lane
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the current-runtime policy probe
+        id: compatibility-probe
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-jpa:test
+          --tests '${{ matrix.test }}'
+          --stacktrace
+      - name: Record what this lane did and did not execute
+        if: always()
+        env:
+          PROBE_OUTCOME: ${{ steps.compatibility-probe.outcome }}
+          TARGET: ${{ matrix.target }}
+          TARGET_COORDINATE: ${{ matrix.coordinate }}
+          REASON: ${{ matrix.reason }}
+        run: |
+          mkdir -p compatibility-evidence
+          {
+            echo "target=${TARGET}"
+            echo "target-coordinate=${TARGET_COORDINATE}"
+            echo "status=NOT_EXECUTABLE"
+            echo "probe-result=${PROBE_OUTCOME}"
+            echo "reason=${REASON}"
+            echo "what-ran=the current runtime's own policy and lane-definition tests"
+            echo "sha=${{ github.sha }}"
+          } > compatibility-evidence/status.properties
+          echo "::notice::${TARGET} compatibility is NOT_EXECUTABLE: ${REASON}"
+      - name: Upload the compatibility status
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: compatibility-status-${{ matrix.id }}
+          path: compatibility-evidence/status.properties
+          retention-days: 30
+          if-no-files-found: error
@@ -0,0 +1,90 @@
+name: jpa-nightly
+
+# The suites that are too slow, too Docker-heavy, or too machine-dependent for a PR, and the middle
+# of the PostgreSQL matrix.
+#
+# The failure-injection lane is the one that matters most and is easiest to lose: it is the only
+# place the commit-ambiguity scenarios run, and they are the only evidence that a lost commit
+# acknowledgement produces completion-unknown rather than a retry.
+
+on:
+  workflow_dispatch:
+  schedule:
+    # 02:30 UTC daily.
+    - cron: '30 2 * * *'
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}
+  cancel-in-progress: false
+
+jobs:
+  jpa-full-matrix:
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    strategy:
+      fail-fast: false
+      matrix:
+        postgresql: ["16", "17", "18"]
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Certify the platform against PostgreSQL ${{ matrix.postgresql }}
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-jpa:jpaPlatformContractTest
+          -Pjpa.matrix.versions=${{ matrix.postgresql }}
+
+          --stacktrace
+
+  jpa-failure-injection:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Reproduce deadlock, serialization, and commit-ambiguity scenarios
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-jpa:jpaPlatformFailureTest
+
+          --stacktrace
+
+  jpa-query-plan-and-security:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the query plan and database security suites
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-jpa:jpaPlatformQueryPlanTest
+          :adapter:outbound:persistence-jpa:jpaPlatformSecurityTest
+
+          --stacktrace
+
+  jpa-pool-pressure:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Verify pool saturation and REQUIRES_NEW connection behaviour
+        working-directory: src
+        # A behaviour contract, not a measurement. This step used to switch assertions off with an
+        # explicit property and call the result a certification, so the only threshold it ever
+        # asserted was that thresholds were not being asserted. What
+        # it checks now — that REQUIRES_NEW needs two connections per concurrent thread, that a
+        # saturated pool reports its pending count, that a caller waits rather than proceeding
+        # without a connection — is true on any runner, so there is nothing to switch off.
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-jpa:jpaPlatformPoolContractTest
+
+          --stacktrace
@@ -26,22 +26,14 @@ jobs:
      JPA_EVIDENCE_TOPOLOGY: postgresql-16-testcontainers-tls-and-fault-matrix
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
+      - uses: ./.github/actions/setup-gradle-java
      - name: Verify the production-profile JPA R2 manifest DAG
        working-directory: src
        run: >-
          ./gradlew
          :adapter:outbound:persistence-jpa:verifyJpaPrimaryFoundationEvidence
          -PjpaEvidenceProfile=r2
-          --no-daemon
+
          --stacktrace
      - name: Retain JPA R2 attempt manifests
        if: always()
@@ -0,0 +1,128 @@
+name: jpa-release
+
+# The release registry is the gate-task source: jpaReleaseQualification reads its blocking gates,
+# while JpaReleaseRenderingTest holds this file's matrix and promotion lists to the registry's Stable
+# majors and verifyJpaReleaseGateTasks resolves each declared task against the real Gradle graph.
+# CI therefore owns release scheduling, not a second JPA gate-task inventory.
+#
+# The matrix below is therefore not free to drift: editing it without editing the registry fails the
+# unit lane.
+
+on:
+  workflow_dispatch:
+  push:
+    tags:
+      - 'v*'
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: false
+
+jobs:
+  # One job per PostgreSQL major, because one job for three majors was one job for one major.
+  #
+  # `-Pjpa.matrix.versions=16,17,18` reached JpaPlatformContractSupport.start(), which started
+  # selectedVersions().get(0) — so twenty-eight integration classes ran against PG16 and nothing
+  # ran against 17 or 18, while docs/jpa/support-matrix.md recorded all three as "full contract
+  # suite, release lane". A JSONB mapping, a Hibernate dialect difference or a Flyway upgrade that
+  # only breaks on 18 shipped with a green release.
+  #
+  # start() now fails closed on a multi-version selection, so the fan-out is not optional: the
+  # matrix is the only way the three majors get covered, and removing a major from it removes the
+  # evidence rather than quietly reusing another major's.
+  jpa-release-gate:
+    runs-on: ubuntu-latest
+    timeout-minutes: 90
+    strategy:
+      fail-fast: false
+      matrix:
+        postgresql: ["16", "17", "18"]
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the JPA database qualification set on PostgreSQL ${{ matrix.postgresql }}
+        working-directory: src
+        run: >-
+          ./gradlew
+          jpaReleaseQualification
+          -Pjpa.matrix.versions=${{ matrix.postgresql }}
+          --stacktrace
+      - name: Record which major this evidence covers
+        if: always()
+        working-directory: src
+        run: |
+          mkdir -p build/jpa-release-evidence
+          {
+            echo "sha=${{ github.sha }}"
+            echo "ref=${{ github.ref }}"
+            echo "postgresql-major=${{ matrix.postgresql }}"
+            echo "task-set=jpa-database-qualification"
+          } > "build/jpa-release-evidence/manifest-${{ matrix.postgresql }}.properties"
+      - name: Upload the release evidence
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: jpa-release-evidence-pg${{ matrix.postgresql }}
+          path: |
+            src/build/jpa-release-evidence/manifest-${{ matrix.postgresql }}.properties
+            src/adapter/outbound/persistence-jpa/build/test-results/**/*.xml
+          retention-days: 30
+          if-no-files-found: error
+
+  # The promotion decision. Three majors' evidence, and all three must come from this SHA — an
+  # aggregate that accepted a re-run artifact from another commit would promote a release on
+  # evidence produced by different code.
+  jpa-release-promotion:
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    needs: jpa-release-gate
+    steps:
+      - name: Download every major's evidence
+        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # actions/download-artifact@v4.3.0
+        with:
+          pattern: jpa-release-evidence-pg*
+          path: evidence
+      - name: Require all three majors, all from this SHA
+        run: |
+          set -euo pipefail
+          missing=0
+          for major in 16 17 18; do
+            manifest=$(find evidence -name "manifest-${major}.properties" -print -quit)
+            if [ -z "${manifest}" ]; then
+              echo "::error::no release evidence for PostgreSQL ${major}"
+              missing=1
+              continue
+            fi
+            sha=$(sed -n 's/^sha=//p' "${manifest}")
+            if [ "${sha}" != "${{ github.sha }}" ]; then
+              echo "::error::PostgreSQL ${major} evidence is from ${sha}, not ${{ github.sha }}"
+              missing=1
+            fi
+          done
+          if [ "${missing}" -ne 0 ]; then
+            echo "::error::the release gate covers three PostgreSQL majors; promotion needs all three"
+            exit 1
+          fi
+          echo "PostgreSQL 16, 17 and 18 evidence all present and all from ${{ github.sha }}."
+
+  jpa-architecture-and-docs:
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Verify architecture boundaries and the support matrix
+        working-directory: src
+        run: >-
+          ./gradlew
+          :verifyJpaReleaseGateTasks
+          :verifyJpaReadinessRegistry
+          :verifyCleanArchitectureDependencies
+          checkstyleMain
+          :app-bootstrap:architectureTest
+          :adapter:outbound:persistence-jpa:test --tests '*JpaReleaseManifestTest'
+
+          --stacktrace
@@ -5,6 +5,8 @@ on:
    paths:
      - "README.md"
      - "src/README.md"
+      - "src/**/README.md"
+      - "src/**/CLAUDE.md"
      - "docs/**/*.md"
      - ".github/**/*.md"
      - ".github/workflows/link-check.yml"
@@ -13,6 +15,8 @@ on:
    paths:
      - "README.md"
      - "src/README.md"
+      - "src/**/README.md"
+      - "src/**/CLAUDE.md"
      - "docs/**/*.md"
      - ".github/**/*.md"
      - ".github/workflows/link-check.yml"
@@ -38,6 +42,8 @@ jobs:
            --root-dir .
            README.md
            src/README.md
+            'src/**/README.md'
+            'src/**/CLAUDE.md'
            'docs/**/*.md'
            '.github/**/*.md'
          fail: true
@@ -0,0 +1,51 @@
+# The messaging platform's broker certification lane.
+#
+# Separate from ci-quality-gates.yml because it needs a container runtime and several minutes of it.
+# The lane deliberately carries no Docker guard: every other container suite in the messaging tree
+# skips with a stated reason when Docker is absent, and a certification lane that skipped would
+# report success for a broker nobody started — which is the exact claim the evidence exists to rule
+# out.
+#
+# The job runs the evidence gate rather than the lane, and the gate depends on the lane. What it
+# proves is not only that the scenarios pass but that the committed manifest
+# (messaging-testkit/src/main/resources/messaging/broker-certification-evidence.jsonl) is what this
+# run produced, so "certified against a live broker" cannot be restored by editing a file.
+name: messaging-certification
+
+on:
+  pull_request:
+    paths:
+      - "src/messaging/**"
+      - ".github/workflows/messaging-certification.yml"
+      # Every Gradle job here installs its toolchain through this composite action, so a change to
+      # it changes what this gate runs.
+      - ".github/actions/setup-gradle-java/action.yml"
+  schedule:
+    - cron: "41 4 * * 3"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+env:
+  # Reuse would hand one scenario the broker another scenario had already faulted.
+  TESTCONTAINERS_REUSE_ENABLE: "false"
+
+jobs:
+  broker-certification:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Certify the Kafka adapter against a real broker
+        working-directory: src
+        # GITHUB_SHA is read by the lane and written into every evidence line, because "certified"
+        # is a claim about one source tree.
+        run: ./gradlew :messaging:messaging-kafka:verifyMessagingCertificationEvidence --stacktrace
+      - name: Publish the certification evidence
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: messaging-broker-certification-evidence
+          path: src/messaging/messaging-kafka/build/messaging-certification/
+          if-no-files-found: warn
@@ -0,0 +1,153 @@
+name: notification-platform
+
+# Verification tiers for the Notification Delivery Platform.
+#
+# The PR tier is deliberately free of any external provider. A gate that depends on a third-party
+# sandbox fails for reasons that have nothing to do with the change under review, and a gate people
+# learn to re-run is not a gate. Real provider smoke tests live in the secret-protected tier, where
+# a failure is an environment signal rather than a merge blocker.
+#
+# Every job that invokes Gradle validates the wrapper first with the repository's pinned action;
+# the wrapper JAR is executable code fetched at build time, so validating it is what keeps a
+# compromised wrapper from turning any workflow run into arbitrary code execution.
+
+on:
+  pull_request:
+    paths:
+      # The filter used to stop at the four notification source trees, so a change to the
+      # composition root, the settings binding, the schema migrations, or the evidence manifest
+      # ran none of this — and those are exactly the surfaces that decide whether the platform
+      # assembles, binds and migrates at all.
+      - 'src/application-core/src/**/notification/**'
+      - 'src/adapter/outbound/notification/**'
+      - 'src/adapter/outbound/persistence-jpa/src/**/notification/**'
+      - 'src/adapter/outbound/persistence-jpa/src/main/resources/db/migration/jpa/notification-platform/**'
+      - 'src/adapter/inbound/web/src/**/notification/**'
+      - 'src/app-bootstrap/src/**/notification/**'
+      - 'src/app-bootstrap/src/main/resources/application*.yml'
+      - 'src/gradle/notification-*.gradle'
+      - 'src/config/architecture/modules.json'
+      - 'src/.env'
+      - 'docs/notification/**'
+      - 'infra/notification/**'
+      - '.github/workflows/notification-platform.yml'
+      # Every Gradle job here installs its toolchain through this composite action, so a change to
+      # it changes what this gate runs.
+      - '.github/actions/setup-gradle-java/action.yml'
+  push:
+    branches: [ main ]
+  schedule:
+    # Nightly: the chaos tier, which is slower and inherently less deterministic than the PR tier.
+    - cron: '0 17 * * *'
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+concurrency:
+  group: notification-platform-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  pr:
+    name: contract (Java 21, no external provider)
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Compile and format check
+        working-directory: src
+        run: ./gradlew :application-core:compileJava :adapter:outbound:notification:compileJava --console=plain
+      - name: Application contracts
+        working-directory: src
+        run: ./gradlew :application-core:test --console=plain
+      - name: Provider contract suite
+        working-directory: src
+        run: ./gradlew :adapter:outbound:notification:test --console=plain
+      - name: Persistence and web
+        working-directory: src
+        run: ./gradlew :adapter:outbound:persistence-jpa:test :adapter:inbound:web:test --console=plain
+      # The PR tier never touched a database, so every claim about migrations, claim atomicity and
+      # lease fencing rested on a fake. Docker is available on this runner; the lane fails closed
+      # when the container cannot start, because a skipped contract reports success for a database
+      # nobody tested.
+      - name: Notification schema and claim contracts (real PostgreSQL)
+        working-directory: src
+        run: ./gradlew :adapter:outbound:persistence-jpa:jpaPlatformContractTest --console=plain
+      - name: Notification migration upgrade (real PostgreSQL)
+        working-directory: src
+        run: ./gradlew :adapter:outbound:persistence-jpa:jpaPlatformMigrationTest --console=plain
+      - name: Architecture gates
+        working-directory: src
+        run: |
+          ./gradlew :verifyCleanArchitectureDependencies --console=plain
+          ./gradlew :app-bootstrap:architectureTest --console=plain
+      - name: Configuration surface
+        working-directory: src
+        run: |
+          ./gradlew :app-bootstrap:verifyEnvKeys :verifyPublicPathSnapshot --console=plain
+          ./gradlew :verifyNotificationApiSurface :verifyNotificationConfiguration --console=plain
+      # A support grade is a promise about production behaviour. This refuses one the pipeline
+      # cannot back — the check that would have caught five channels reading "Stable" while no
+      # request had ever left the process.
+      - name: Evidence manifest
+        working-directory: src
+        run: ./gradlew :verifyNotificationEvidence --console=plain
+      - name: Static analysis
+        working-directory: src
+        run: ./gradlew :adapter:outbound:notification:check -x test --console=plain
+
+  nightly-chaos:
+    name: chaos (ambiguity, restart recovery, callback burst)
+    if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      # This job is named for ambiguity, restart recovery and callback burst. It used to run a
+      # unit-test filter and then `test` — neither of which restarts anything or bursts anything —
+      # so the job name was the only place those three properties existed.
+      - name: Ambiguity and fault harness
+        working-directory: src
+        run: ./gradlew :adapter:outbound:notification:test --tests '*ChaosSecurity*' --tests '*CrossProviderContractSuite*' --console=plain
+      - name: Concurrency and rotation races
+        working-directory: src
+        run: ./gradlew :adapter:outbound:notification:test --tests '*ConcurrencyTest' --tests '*ProviderRuntimeStateTest' --console=plain
+      - name: Restart recovery and lease fencing (real PostgreSQL)
+        working-directory: src
+        run: ./gradlew :adapter:outbound:persistence-jpa:jpaPlatformContractTest :adapter:outbound:persistence-jpa:jpaPlatformFailureTest --console=plain
+      - name: Notification regression suite
+        working-directory: src
+        run: >-
+          ./gradlew
+          :application-core:test
+          :adapter:outbound:notification:test
+          :adapter:outbound:persistence-jpa:test
+          :adapter:inbound:web:test
+          --console=plain
+      # A filter that matches nothing passes. Each --tests filter above names a class that exists
+      # today; if one is renamed the job must fail rather than quietly stop covering it.
+      - name: Every named suite actually ran
+        working-directory: src
+        run: |
+          set -euo pipefail
+          for suite in ChaosSecurity CrossProviderContractSuite ConcurrencyTest ProviderRuntimeStateTest; do
+            if ! find . -path '*/build/test-results/*' -name "*${suite}*.xml" | grep -q .; then
+              echo "no test results for ${suite}: the filter matched nothing and the job passed vacuously" >&2
+              exit 1
+            fi
+          done
+
+  # There is no provider-sandbox job. It ran only on workflow_dispatch and could not succeed by
+  # any path: with no credentials its first step exit 1-ed, and with credentials the only test it
+  # ran was ProviderSandboxSmokeTest, whose body is an unconditional fail() saying a real sandbox
+  # call is not implemented. Its credential check read secrets.NOTIFICATION_SANDBOX_CREDENTIALS,
+  # which nothing in this repository consumes — the test reads NOTIFICATION_SANDBOX_ENABLED — so
+  # any non-empty string satisfied it and was then dropped.
+  #
+  # The unimplemented state is still stated in two places that do not depend on a workflow:
+  # ProviderSandboxSmokeTest itself, and the unsatisfied provider-wire-qualified claim in
+  # docs/notification/evidence-manifest.json, which verifyNotificationEvidence enforces inside
+  # check. When a real sandbox call is implemented, the job comes back with it.
@@ -11,6 +11,17 @@ on:
        required: true
        default: false
        type: boolean
+      # AwsS3DirectTransferQualificationTest requires a second, separate authority
+      # (OBJECT_STORAGE_AWS_DIRECT_MUTATION_ENABLED) before the direct-transfer lane may run, and
+      # the job never supplied it. objectStorageAwsQualificationTest is a strict qualification task
+      # that requires both of its classes, so the lane could not be run to a pass from any input:
+      # dispatching it always failed on the missing variable. The authority now exists as its own
+      # input rather than as a constant, which is what "separate" was supposed to mean.
+      run_protected_aws_direct_mutation:
+        description: Also authorize the direct-transfer mutation lane against the sandbox bucket
+        required: true
+        default: false
+        type: boolean

 permissions:
  contents: read
@@ -19,78 +30,41 @@ env:
  TESTCONTAINERS_REUSE_ENABLE: "false"

 jobs:
-  poster-image-v7-migration:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
-      - name: Run non-skipping Poster image V7 migration qualification
-        working-directory: src
-        run: ./gradlew :sample-portfolio:posterImageMigrationTest --no-daemon --stacktrace
-
  minio-managed-contract:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
+      - uses: ./.github/actions/setup-gradle-java
      - name: Run exact-release MinIO managed contract
        working-directory: src
-        run: ./gradlew :adapter:outbound:objectstorage:objectStorageMinioContractTest --no-daemon --stacktrace
+        run: ./gradlew :adapter:outbound:objectstorage:objectStorageMinioContractTest --stacktrace

  minio-managed-fault:
    if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
+      - uses: ./.github/actions/setup-gradle-java
      - name: Run digest-pinned MinIO and Toxiproxy fault contract
        working-directory: src
-        run: ./gradlew :adapter:outbound:objectstorage:objectStorageMinioFaultTest --no-daemon --stacktrace
+        run: ./gradlew :adapter:outbound:objectstorage:objectStorageMinioFaultTest --stacktrace

  aws-managed-common-subset:
-    if: github.event_name == 'workflow_dispatch' && inputs.run_protected_aws
+    if: >-
+      github.event_name == 'workflow_dispatch'
+      && inputs.run_protected_aws
+      && inputs.run_protected_aws_direct_mutation
    environment: object-storage-aws-qualification
    runs-on: ubuntu-latest
    env:
      OBJECT_STORAGE_AWS_QUALIFICATION_ENABLED: "true"
+      OBJECT_STORAGE_AWS_DIRECT_MUTATION_ENABLED: ${{ inputs.run_protected_aws_direct_mutation }}
      OBJECT_STORAGE_AWS_BUCKET: ${{ secrets.OBJECT_STORAGE_AWS_BUCKET }}
      OBJECT_STORAGE_AWS_REGION: ${{ secrets.OBJECT_STORAGE_AWS_REGION }}
      OBJECT_STORAGE_AWS_EXPECTED_OWNER: ${{ secrets.OBJECT_STORAGE_AWS_EXPECTED_OWNER }}
    steps:
      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-          cache-dependency-path: |
-            src/**/*.gradle
-            src/**/gradle-wrapper.properties
-            src/**/gradle.lockfile
+      - uses: ./.github/actions/setup-gradle-java
      - name: Run protected AWS common-subset qualification
        working-directory: src
-        run: ./gradlew :adapter:outbound:objectstorage:objectStorageAwsQualificationTest --no-daemon --stacktrace
+        run: ./gradlew :adapter:outbound:objectstorage:objectStorageAwsQualificationTest --stacktrace
@@ -0,0 +1,289 @@
+name: pr-adapters
+
+# Stage 1, the adapter half: the lanes a pull request must clear that `ci-quality-gates.yml` cannot
+# reach.
+#
+# It replaces web-pr.yml, websocket-pr.yml, httpclient-contract.yml and jpa-pr.yml, which were four
+# files split by module rather than by stage. Splitting by module is what made the duplication
+# invisible: each file opened with its own "unit and architecture" job running
+# `:<leaf>:test verifyCleanArchitectureDependencies`, and all four of those were already inside the
+# root `check` that ci-quality-gates.yml runs on every pull request with no path filter. Four jobs,
+# four runners, four Gradle configurations, zero additional coverage. They are gone; what is left
+# here is only what `check` does not run.
+#
+# What `check` does not run, and therefore what this file is for:
+#   * lanes with their own source set — a second servlet container, a real Nginx, Reactor Netty;
+#   * lanes selected by a tag that `test` excludes — the cross-stack parity recording comparison;
+#   * lanes parameterised per run — one PostgreSQL major per job, one HTTP transport per job.
+# Each of those genuinely cannot run inside `check`, which is the test for whether a job belongs
+# here at all.
+#
+# Path filtering is per job rather than per workflow. The four files it replaces each carried an
+# `on.pull_request.paths` list, so the whole file was skipped or run as a unit; a change touching
+# web and JPA started two workflows and a change touching neither still started none. Here one
+# `changes` job computes the answer once from the pull request's own diff and every lane reads it.
+# The filter is a plain `git diff` rather than a filter action: this repository pins every action by
+# commit SHA and adding a third-party action to compute a boolean is a supply-chain decision, not a
+# convenience.
+
+on:
+  pull_request:
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  # One diff, read once. `workflow_dispatch` answers "everything changed", because a manual run is
+  # somebody asking for the lanes and there is no base ref to compare against.
+  changes:
+    runs-on: ubuntu-latest
+    timeout-minutes: 10
+    outputs:
+      web: ${{ steps.filter.outputs.web }}
+      websocket: ${{ steps.filter.outputs.websocket }}
+      httpclient: ${{ steps.filter.outputs.httpclient }}
+      jpa: ${{ steps.filter.outputs.jpa }}
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+        with:
+          # Both endpoints of the pull request's diff have to be present locally; the default
+          # shallow fetch has neither the base commit nor the merge base.
+          fetch-depth: 0
+      - name: Decide which adapter lanes this diff can affect
+        id: filter
+        env:
+          BASE_SHA: ${{ github.event.pull_request.base.sha }}
+          HEAD_SHA: ${{ github.event.pull_request.head.sha }}
+        run: |
+          set -euo pipefail
+          if [ "${GITHUB_EVENT_NAME}" != 'pull_request' ]; then
+            changed='ALL'
+          else
+            if [ -z "${BASE_SHA}" ] || [ -z "${HEAD_SHA}" ]; then
+              echo "::error::pull request diff endpoints are missing; refusing to report no lanes"
+              exit 1
+            fi
+            changed="$(git diff --name-only "${BASE_SHA}" "${HEAD_SHA}")"
+          fi
+          # Fail closed rather than reporting "nothing changed": an empty diff on a pull request
+          # means the comparison did not work, and a filter that answers false on a broken
+          # comparison silently turns every lane below off.
+          if [ "${changed}" != 'ALL' ] && [ -z "${changed}" ]; then
+            echo "::error::the pull request diff is empty; the comparison did not run"
+            exit 1
+          fi
+          printf 'changed files:\n%s\n' "${changed}"
+          emit() {
+            lane="$1"
+            shift
+            if [ "${changed}" = 'ALL' ]; then
+              printf '%s=true\n' "${lane}" >> "${GITHUB_OUTPUT}"
+              printf 'lane %s: true (manual run)\n' "${lane}"
+              return 0
+            fi
+            for pattern in "$@"; do
+              if printf '%s\n' "${changed}" | grep -qE -- "${pattern}"; then
+                printf '%s=true\n' "${lane}" >> "${GITHUB_OUTPUT}"
+                printf 'lane %s: true (%s)\n' "${lane}" "${pattern}"
+                return 0
+              fi
+            done
+            printf '%s=false\n' "${lane}" >> "${GITHUB_OUTPUT}"
+            printf 'lane %s: false\n' "${lane}"
+          }
+          # This workflow and the composite action every lane below uses are in every lane's path
+          # set: a change to either changes what the lanes do, and a gate that does not re-run when
+          # its own definition changes is a gate nobody has seen run in its current form.
+          common='^\.github/workflows/pr-adapters\.yml$|^\.github/actions/'
+          emit web \
+            '^src/adapter/inbound/web/' \
+            '^src/application-core/src/.*/operation/' \
+            '^src/application-core/src/.*/idempotency/' \
+            '^src/adapter/outbound/persistence-jpa/src/.*/operation/' \
+            '^docs/web/' \
+            "${common}"
+          emit websocket \
+            '^src/adapter/inbound/websocket/' \
+            '^docs/websocket/' \
+            "${common}"
+          emit httpclient \
+            '^src/adapter/outbound/httpclient/' \
+            '^src/app-bootstrap/src/.*/httpclient/' \
+            '^docs/httpclient/' \
+            '^scripts/verify-httpclient-docs\.py$' \
+            "${common}"
+          emit jpa \
+            '^src/adapter/outbound/persistence-jpa/' \
+            '^src/app-bootstrap/src/.*/jpa/' \
+            '^src/config/architecture/modules\.json$' \
+            '^docs/jpa/' \
+            '^docs/adr/ADR-JPA-' \
+            '^infra/jpa/' \
+            "${common}"
+
+  # The parity gate depends on all three recording lanes and fails when one is missing, so it runs
+  # them itself rather than trusting a previous job to have left the recordings behind. Its tag is
+  # excluded from `test`, which is why `check` cannot cover it.
+  web-cross-stack-parity:
+    needs: changes
+    if: needs.changes.outputs.web == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 40
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Compare the wire contract across Tomcat, Jetty and Reactor Netty
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:web:webCrossStackParityTest
+
+          --stacktrace
+      - name: Publish the parity recordings
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: web-contract-parity
+          path: src/adapter/inbound/web/build/web-contract-parity/
+          if-no-files-found: error
+
+  # Docker-gated, and the lane fails rather than skipping when the runtime is missing. A proxy
+  # contract that quietly passes without a proxy has been certifying nothing since whenever the
+  # container runtime last broke.
+  web-nginx-proxy-contract:
+    needs: changes
+    if: needs.changes.outputs.web == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the proxy, prefix and spoofing contract behind a real Nginx
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:web:webNginxProxyTest
+
+          --stacktrace
+
+  websocket-container-matrix:
+    needs: changes
+    if: needs.changes.outputs.websocket == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the runtime contract on the second servlet container
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:websocket:websocketJettyTest
+
+          --stacktrace
+
+  # Docker-gated, and the lane fails rather than skipping. Upgrade handling is the single most
+  # common WebSocket deployment failure and it is invisible from either side alone.
+  websocket-nginx-contract:
+    needs: changes
+    if: needs.changes.outputs.websocket == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the upgrade and forwarded-header contract behind a real Nginx
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:websocket:websocketNginxTest
+
+          --stacktrace
+
+  # One transport per job, so a transport that stops satisfying the shared contract fails on its own
+  # row instead of disappearing into an aggregate run. `check` runs this lane once, unparameterised.
+  httpclient-stable-contract:
+    needs: changes
+    if: needs.changes.outputs.httpclient == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    strategy:
+      fail-fast: false
+      matrix:
+        transport: [apache, jdk, reactor]
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Certify one transport against the shared contract
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:httpclient:httpClientStableContractTest
+          -Phttpclient.contract.transports=${{ matrix.transport }}
+
+          --stacktrace
+
+  # Only the Spring 7.0 lane. httpClientSecurityTest, httpClientBlockHoundTest and
+  # spring62ApiSurfaceScan used to run here too; all three are `dependsOn` of this leaf's `check`
+  # (src/adapter/outbound/httpclient/build.gradle), so ci-quality-gates.yml already ran them on the
+  # same pull request. spring70CompatibilityTest is deliberately outside `check` and is what is left.
+  httpclient-security-and-compatibility:
+    needs: changes
+    if: needs.changes.outputs.httpclient == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the next-major Spring compatibility lane
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:httpclient:spring70CompatibilityTest
+
+          --stacktrace
+
+  # 16 and 18 — the ends of the Stable matrix. 17 runs in the integration stage. What this does not
+  # do is skip the container lane on a runner without Docker: PostgreSqlContainerFactory throws,
+  # because a skipped contract reports success for a database nobody tested.
+  jpa-postgresql-contract:
+    needs: changes
+    if: needs.changes.outputs.jpa == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    strategy:
+      fail-fast: false
+      matrix:
+        postgresql: ["16", "18"]
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Certify the platform against PostgreSQL ${{ matrix.postgresql }}
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-jpa:jpaPlatformContractTest
+          -Pjpa.matrix.versions=${{ matrix.postgresql }}
+
+          --stacktrace
+
+  jpa-migration-smoke:
+    needs: changes
+    if: needs.changes.outputs.jpa == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the migration upgrade smoke scenarios
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:outbound:persistence-jpa:jpaPlatformMigrationTest
+
+          --stacktrace
@@ -1,375 +0,0 @@
-name: redis-production-readiness
-
-on:
-  schedule:
-    - cron: "23 18 * * *"
-  workflow_dispatch:
-  push:
-    tags:
-      - "v*-rc.*"
-
-permissions:
-  contents: read
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.ref }}
-  cancel-in-progress: false
-
-jobs:
-  resolve-redis-readiness:
-    runs-on: ubuntu-latest
-    outputs:
-      selected: ${{ steps.resolve.outputs.selected }}
-      candidates: ${{ steps.resolve.outputs.candidates }}
-      selected_count: ${{ steps.resolve.outputs.selected_count }}
-      sentinel_required: ${{ steps.resolve.outputs.sentinel_required }}
-      cluster_required: ${{ steps.resolve.outputs.cluster_required }}
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - name: Generate the strict checked-in readiness control artifact
-        working-directory: src
-        run: ./gradlew writeRedisCiMatrix --no-daemon --stacktrace
-      - id: resolve
-        name: Transport the generated matrix to job outputs
-        shell: python
-        run: |
-          import json
-          import os
-          from pathlib import Path
-
-          matrix_path = Path(
-              "src/build/redis-evidence/control/redis-readiness-matrix.json"
-          )
-          matrix = json.loads(matrix_path.read_text(encoding="utf-8"))
-          if matrix["releaseQualification"] != "NOT_CLAIMED":
-              raise SystemExit("resolver control artifact must not claim release qualification")
-          with Path(os.environ["GITHUB_OUTPUT"]).open("a", encoding="utf-8") as stream:
-              stream.write(
-                  "selected="
-                  + json.dumps(matrix["selected"], separators=(",", ":"))
-                  + "\n"
-              )
-              stream.write(
-                  "candidates="
-                  + json.dumps(
-                      matrix["implementedCandidates"], separators=(",", ":")
-                  )
-                  + "\n"
-              )
-              stream.write(f"selected_count={matrix['selectedCount']}\n")
-              stream.write(
-                  "sentinel_required="
-                  + str(matrix["topologyJobs"]["sentinel"]).lower()
-                  + "\n"
-              )
-              stream.write(
-                  "cluster_required="
-                  + str(matrix["topologyJobs"]["cluster"]).lower()
-                  + "\n"
-              )
-      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-readiness-control
-          path: src/build/redis-evidence/control
-          if-no-files-found: error
-          retention-days: 30
-
-  redis-security:
-    needs: resolve-redis-readiness
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - id: redis-tests
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:redisSecurityTest --no-daemon --stacktrace
-      - id: redis-evidence-sanitizer
-        if: always()
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:verifyRedisEvidenceArtifactsForUpload --no-daemon --stacktrace
-      - if: ${{ always() && steps.redis-evidence-sanitizer.outcome == 'success' }}
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-security-evidence
-          path: src/adapter/outbound/cache-redis/build/redis-evidence
-          if-no-files-found: error
-          retention-days: 14
-
-  redis-sentinel:
-    needs: resolve-redis-readiness
-    if: ${{ needs.resolve-redis-readiness.outputs.sentinel_required == 'true' }}
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - id: redis-tests
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:redisSentinelTest --no-daemon --stacktrace
-      - id: redis-evidence-sanitizer
-        if: always()
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:verifyRedisEvidenceArtifactsForUpload --no-daemon --stacktrace
-      - if: ${{ always() && steps.redis-evidence-sanitizer.outcome == 'success' }}
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-sentinel-evidence
-          path: src/adapter/outbound/cache-redis/build/redis-evidence
-          if-no-files-found: error
-          retention-days: 14
-
-  redis-cluster:
-    needs: resolve-redis-readiness
-    if: ${{ needs.resolve-redis-readiness.outputs.cluster_required == 'true' }}
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - id: redis-tests
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:redisClusterTest --no-daemon --stacktrace
-      - id: redis-evidence-sanitizer
-        if: always()
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:verifyRedisEvidenceArtifactsForUpload --no-daemon --stacktrace
-      - if: ${{ always() && steps.redis-evidence-sanitizer.outcome == 'success' }}
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-cluster-evidence
-          path: src/adapter/outbound/cache-redis/build/redis-evidence
-          if-no-files-found: error
-          retention-days: 14
-
-  redis-fault:
-    needs: resolve-redis-readiness
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - id: redis-tests
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:redisFaultTest --no-daemon --stacktrace
-      - id: redis-evidence-sanitizer
-        if: always()
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:verifyRedisEvidenceArtifactsForUpload --no-daemon --stacktrace
-      - if: ${{ always() && steps.redis-evidence-sanitizer.outcome == 'success' }}
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-fault-evidence
-          path: src/adapter/outbound/cache-redis/build/redis-evidence
-          if-no-files-found: error
-          retention-days: 14
-
-  redis-compatibility:
-    needs: resolve-redis-readiness
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - id: redis-tests
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:redisCompatibilityTest --no-daemon --stacktrace
-      - id: redis-evidence-sanitizer
-        if: always()
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:verifyRedisEvidenceArtifactsForUpload --no-daemon --stacktrace
-      - if: ${{ always() && steps.redis-evidence-sanitizer.outcome == 'success' }}
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-compatibility-evidence
-          path: src/adapter/outbound/cache-redis/build/redis-evidence
-          if-no-files-found: error
-          retention-days: 14
-
-  selected-card-readiness:
-    needs: resolve-redis-readiness
-    if: ${{ needs.resolve-redis-readiness.outputs.selected_count != '0' }}
-    strategy:
-      fail-fast: false
-      matrix:
-        include: ${{ fromJSON(needs.resolve-redis-readiness.outputs.selected) }}
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - id: redis-tests
-        working-directory: src
-        run: ./gradlew ${{ matrix.readinessTask }} --no-daemon --stacktrace
-      - id: redis-evidence-sanitizer
-        if: always()
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:verifyRedisEvidenceArtifactsForUpload --no-daemon --stacktrace
-      - if: ${{ always() && steps.redis-evidence-sanitizer.outcome == 'success' }}
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-selected-${{ matrix.cardId }}
-          path: src/adapter/outbound/cache-redis/build/redis-evidence
-          if-no-files-found: error
-          retention-days: 30
-
-  redis-all-candidates:
-    needs: resolve-redis-readiness
-    if: ${{ github.event_name == 'schedule' || github.event_name == 'workflow_dispatch' }}
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - id: redis-tests
-        working-directory: src
-        run: ./gradlew redisAllImplementedCandidates --no-daemon --stacktrace
-      - id: redis-evidence-sanitizer
-        if: always()
-        working-directory: src
-        run: ./gradlew :adapter:outbound:cache-redis:verifyRedisEvidenceArtifactsForUpload --no-daemon --stacktrace
-      - if: ${{ always() && steps.redis-evidence-sanitizer.outcome == 'success' }}
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4
-        with:
-          name: redis-all-candidates-evidence
-          path: src/adapter/outbound/cache-redis/build/redis-evidence
-          if-no-files-found: error
-          retention-days: 14
-
-  redis-production-readiness:
-    needs:
-      - resolve-redis-readiness
-      - selected-card-readiness
-    if: ${{ always() && needs.resolve-redis-readiness.result == 'success' }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Require the exact selected matrix result
-        shell: python
-        env:
-          SELECTED_COUNT: ${{ needs.resolve-redis-readiness.outputs.selected_count }}
-          SELECTED_JOB_RESULT: ${{ needs.selected-card-readiness.result }}
-        run: |
-          import os
-
-          selected_count_text = os.environ["SELECTED_COUNT"]
-          selected_job_result = os.environ["SELECTED_JOB_RESULT"]
-          if not selected_count_text.isdecimal():
-              raise SystemExit("selected_count must be a non-negative integer")
-          selected_count = int(selected_count_text)
-          expected_result = "skipped" if selected_count == 0 else "success"
-          if selected_job_result != expected_result:
-              raise SystemExit(
-                  "selected-card-readiness result mismatch: "
-                  f"count={selected_count}, expected={expected_result}, "
-                  f"actual={selected_job_result}"
-              )
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
-      - uses: actions/setup-java@c5195efecf7bdfc987ee8bae7a71cb8b11521c00 # actions/setup-java@v4.7.1
-        with:
-          distribution: temurin
-          java-version: "21.0.11+10"
-          cache: gradle
-      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # actions/download-artifact@v4.3.0
-        with:
-          name: redis-readiness-control
-          path: ${{ runner.temp }}/redis-readiness/control
-      - if: >-
-          ${{
-            needs.resolve-redis-readiness.outputs.selected_count != '0'
-            && needs.selected-card-readiness.result == 'success'
-          }}
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # actions/download-artifact@v4.3.0
-        with:
-          pattern: redis-selected-*
-          path: ${{ runner.temp }}/redis-readiness/selected
-      - name: Record the downloaded selected artifact inventory
-        shell: python
-        env:
-          GITHUB_RUN_ID: ${{ github.run_id }}
-          SELECTED_JSON: ${{ needs.resolve-redis-readiness.outputs.selected }}
-          SELECTED_COUNT: ${{ needs.resolve-redis-readiness.outputs.selected_count }}
-          SELECTED_JOB_RESULT: ${{ needs.selected-card-readiness.result }}
-          REDIS_SELECTED_DIRECTORY: ${{ runner.temp }}/redis-readiness/selected
-          REDIS_CI_RESULT_FILE: ${{ runner.temp }}/redis-readiness/redis-ci-result.json
-        run: |
-          import json
-          import os
-          from pathlib import Path
-
-          selected = json.loads(os.environ["SELECTED_JSON"])
-          if not isinstance(selected, list):
-              raise SystemExit("selected matrix must be a JSON array")
-          expected_names = sorted(
-              "redis-selected-" + entry["cardId"] for entry in selected
-          )
-          if len(expected_names) != int(os.environ["SELECTED_COUNT"]):
-              raise SystemExit("selected_count does not match the selected matrix")
-          if len(expected_names) != len(set(expected_names)):
-              raise SystemExit("selected matrix contains duplicate artifact names")
-
-          selected_directory = Path(os.environ["REDIS_SELECTED_DIRECTORY"])
-          actual_names = (
-              sorted(path.name for path in selected_directory.iterdir() if path.is_dir())
-              if selected_directory.is_dir()
-              else []
-          )
-          if actual_names != expected_names:
-              raise SystemExit(
-                  "downloaded selected artifact inventory mismatch: "
-                  f"expected={expected_names}, actual={actual_names}"
-              )
-
-          result = {
-              "schemaVersion": 1,
-              "runId": os.environ["GITHUB_RUN_ID"],
-              "selectedCount": len(expected_names),
-              "selectedJobResult": os.environ["SELECTED_JOB_RESULT"],
-              "selectedArtifactNames": actual_names,
-          }
-          result_path = Path(os.environ["REDIS_CI_RESULT_FILE"])
-          result_path.parent.mkdir(parents=True, exist_ok=True)
-          result_path.write_text(
-              json.dumps(result, sort_keys=True, separators=(",", ":")) + "\n",
-              encoding="utf-8",
-          )
-      - if: ${{ needs.resolve-redis-readiness.outputs.selected_count == '0' }}
-        working-directory: src
-        run: >-
-          ./gradlew verifyRedisSelectedEvidenceArtifacts redisProductionReadiness
-          -PredisControlDirectory=${{ runner.temp }}/redis-readiness/control
-          -PredisCiResultFile=${{ runner.temp }}/redis-readiness/redis-ci-result.json
-          --no-daemon --stacktrace
-      - if: ${{ needs.resolve-redis-readiness.outputs.selected_count != '0' }}
-        working-directory: src
-        run: >-
-          ./gradlew verifyRedisSelectedEvidenceArtifacts redisProductionReadiness
-          -PredisControlDirectory=${{ runner.temp }}/redis-readiness/control
-          -PredisEvidenceDirectory=${{ runner.temp }}/redis-readiness/selected
-          -PredisCiResultFile=${{ runner.temp }}/redis-readiness/redis-ci-result.json
-          --no-daemon --stacktrace
@@ -0,0 +1,193 @@
+# Redis SDK topology evidence.
+#
+# The lanes in infra/redis-sdk answer what the deterministic in-memory gateway cannot — Sentinel
+# promotion behaviour, Cluster redirects, ACL coverage. docs/redis/support-matrix.md records which
+# lane produced which evidence, and RedisSupportMatrixTest refuses an evidence claim that does not
+# name the test class behind it.
+#
+# Three cadences, because the cost and the question differ:
+#
+#   pull_request  standalone only, current supported version. The cheapest lane that can still
+#                 catch "this change cannot talk to a real Redis at all". A PR gate that starts
+#                 three topologies is a PR gate people learn to ignore.
+#   schedule      the full supported-version x topology matrix, nightly. This is where Sentinel
+#                 promotion and Cluster redirect evidence comes from.
+#   workflow_dispatch  one lane on demand, for reproducing a specific failure.
+#
+# A release candidate uses the nightly matrix run for its tag: `release-candidate` selects the full
+# matrix on demand so an RC does not have to wait for the next scheduled run.
+#
+# Each lane has its own endpoint. A sentinel is not a data node and a cluster node is not the whole
+# cluster, so the address, port, and (for Sentinel) the monitored primary's name are per-lane rather
+# than one hardcoded 6379 that happens to be right for standalone only.
+#
+# The Gradle task is fail-closed on its own account: an unknown mode, a missing endpoint, a lane
+# with no tagged test class, and a run that executed zero tests are all errors. This workflow does
+# not need to re-check those, but it does have to keep the evidence, which is why every run uploads
+# the JUnit XML together with the commit SHA, the server version and the resolved image digest. An
+# evidence artifact that cannot say which image produced it is not evidence.
+name: redis-sdk-topology
+
+on:
+  pull_request:
+    paths:
+      - "src/adapter/outbound/cache-redis/**"
+      - "infra/redis-sdk/**"
+      - ".github/workflows/redis-sdk-topology.yml"
+      # Every Gradle job here installs its toolchain through this composite action, so a change to
+      # it changes what this gate runs.
+      - ".github/actions/setup-gradle-java/action.yml"
+  schedule:
+    # 02:30 UTC daily. Nightly, not hourly: the matrix starts real servers.
+    - cron: "30 2 * * *"
+  workflow_dispatch:
+    inputs:
+      topology:
+        description: standalone, sentinel, cluster, tls, or release-candidate for the full matrix
+        required: true
+        default: standalone
+        type: choice
+        options: [standalone, sentinel, cluster, tls, release-candidate]
+      redis_version:
+        description: server version tag
+        required: true
+        default: "7.4"
+        type: string
+
+permissions:
+  contents: read
+
+jobs:
+  # The matrix is computed rather than duplicated per trigger, so adding a supported version is one
+  # edit and no trigger can silently keep testing an old set.
+  lanes:
+    runs-on: ubuntu-latest
+    outputs:
+      matrix: ${{ steps.select.outputs.matrix }}
+    steps:
+      - id: select
+        run: |
+          set -euo pipefail
+          case "${{ github.event_name }}" in
+            pull_request)
+              matrix='{"include":[{"topology":"standalone","redis_version":"7.4"}]}'
+              ;;
+            schedule)
+              matrix='{"include":[
+                {"topology":"standalone","redis_version":"7.2"},
+                {"topology":"standalone","redis_version":"7.4"},
+                {"topology":"standalone","redis_version":"8.2"},
+                {"topology":"sentinel","redis_version":"7.2"},
+                {"topology":"sentinel","redis_version":"7.4"},
+                {"topology":"sentinel","redis_version":"8.2"},
+                {"topology":"cluster","redis_version":"7.2"},
+                {"topology":"cluster","redis_version":"7.4"},
+                {"topology":"cluster","redis_version":"8.2"},
+                {"topology":"tls","redis_version":"7.4"},
+                {"topology":"tls","redis_version":"8.2"}]}'
+              ;;
+            *)
+              if [ "${{ inputs.topology }}" = "release-candidate" ]; then
+                matrix='{"include":[
+                  {"topology":"standalone","redis_version":"7.2"},
+                  {"topology":"standalone","redis_version":"7.4"},
+                  {"topology":"standalone","redis_version":"8.2"},
+                  {"topology":"sentinel","redis_version":"7.2"},
+                  {"topology":"sentinel","redis_version":"7.4"},
+                  {"topology":"sentinel","redis_version":"8.2"},
+                  {"topology":"cluster","redis_version":"7.2"},
+                  {"topology":"cluster","redis_version":"7.4"},
+                  {"topology":"cluster","redis_version":"8.2"},
+                  {"topology":"tls","redis_version":"7.4"},
+                  {"topology":"tls","redis_version":"8.2"}]}'
+              else
+                matrix='{"include":[{"topology":"${{ inputs.topology }}","redis_version":"${{ inputs.redis_version }}"}]}'
+              fi
+              ;;
+          esac
+          printf 'matrix=%s\n' "$(printf '%s' "$matrix" | tr -d '\n ')" >> "$GITHUB_OUTPUT"
+
+  topology-evidence:
+    needs: lanes
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    strategy:
+      fail-fast: false
+      matrix: ${{ fromJson(needs.lanes.outputs.matrix) }}
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Start the topology
+        env:
+          REDIS_VERSION: ${{ matrix.redis_version }}
+        run: docker compose -f "infra/redis-sdk/${{ matrix.topology }}/compose.yml" up -d --wait
+      - name: Record the image digest
+        id: image
+        run: |
+          set -euo pipefail
+          # The tag says 7.4; the digest says which 7.4. Evidence that names only the tag cannot be
+          # reproduced once the tag moves.
+          #
+          # This used to end in `|| echo 'unresolved'`, which absorbed the failure that `set -e` was
+          # there to catch: the manifest below recorded `image_digest=unresolved`, the upload
+          # satisfied `if-no-files-found: error`, and the lane went green holding exactly the
+          # artifact this workflow's header calls "not evidence". Compose pulls the image in the
+          # step before this one, so RepoDigests is populated; if it is not, the run has nothing to
+          # certify and says so.
+          digest="$(docker image inspect --format '{{index .RepoDigests 0}}' \
+            "redis:${{ matrix.redis_version }}")"
+          if [[ -z "$digest" ]]; then
+            echo "::error::no repository digest for redis:${{ matrix.redis_version }}; this run cannot say which image produced its evidence"
+            exit 1
+          fi
+          printf 'digest=%s\n' "$digest" >> "$GITHUB_OUTPUT"
+      - name: Run the topology contracts
+        working-directory: src
+        run: |
+          set -euo pipefail
+          case '${{ matrix.topology }}' in
+            standalone) port=6379; extra='' ;;
+            sentinel)   port=27010; extra='-Predis.topology.master=skeleton' ;;
+            cluster)    port=7100; extra='' ;;
+            # The TLS lane's CA is generated at start-up, so the trust material is extracted from
+            # the lane rather than checked in. A checked-in key is a secret in the repository
+            # however loudly the file is named "test".
+            tls)
+              port=6390
+              docker compose -f ../infra/redis-sdk/tls/compose.yml cp redis:/tls/ca.crt "$RUNNER_TEMP/redis-lane-ca.pem"
+              extra="-Predis.topology.trust-material=$RUNNER_TEMP/redis-lane-ca.pem"
+              ;;
+            *) echo "unknown topology"; exit 1 ;;
+          esac
+          ./gradlew :adapter:outbound:cache-redis:redisTopologyTest --console=plain \
+            -Predis.topology.host=localhost \
+            -Predis.topology.port="$port" \
+            -Predis.topology.mode='${{ matrix.topology }}' \
+            $extra
+      - name: Write the evidence manifest
+        if: always()
+        run: |
+          set -euo pipefail
+          out=src/adapter/outbound/cache-redis/build/test-results/redisTopologyTest
+          mkdir -p "$out"
+          cat > "$out/evidence-manifest.txt" <<MANIFEST
+          commit=${{ github.sha }}
+          workflow_run=${{ github.run_id }}
+          trigger=${{ github.event_name }}
+          topology=${{ matrix.topology }}
+          redis_version=${{ matrix.redis_version }}
+          image_digest=${{ steps.image.outputs.digest }}
+          MANIFEST
+      - name: Preserve the evidence
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: redis-topology-${{ matrix.topology }}-${{ matrix.redis_version }}
+          path: |
+            src/adapter/outbound/cache-redis/build/test-results/redisTopologyTest/**
+            src/adapter/outbound/cache-redis/build/reports/tests/redisTopologyTest/**
+          if-no-files-found: error
+          retention-days: 90
+      - name: Stop the topology
+        if: always()
+        run: docker compose -f "infra/redis-sdk/${{ matrix.topology }}/compose.yml" down -v
@@ -0,0 +1,466 @@
+name: release
+
+# Stage 3: produce a deployable artifact.
+#
+# One workflow, because there is one deployable unit. `app-bootstrap` is the composition root and
+# the only thing a cluster runs; the adapters are leaves of that artifact, not independently
+# shippable services. Eight files used to answer a release tag — web-release, web-advanced-release,
+# websocket-release, httpclient-release, container-release, and the three that still have to live
+# apart (see below) — and between them they ran `verifyCleanArchitectureDependencies` six times and
+# `:app-bootstrap:test` four times for one release, on separate runners, with no job in any of them
+# able to wait on a job in another.
+#
+# Tag scheme: `v*` only. The adapter-scoped patterns (`web-v*`, `websocket-v*`, `fileserver-v*`) are
+# gone. They were the namespace-split bug: tagging `v1.2.3` and tagging `web-v1.2.3` ran different
+# sets of gates, so a release could choose which gate it cleared, and the adapter-scoped half could
+# not build an image because there is no per-adapter image to build.
+#
+# Two release workflows still stand outside this file, both for a mechanical reason rather than a
+# design one:
+#   * jpa-release.yml — JpaReleaseRenderingTest reads that exact path and holds its PostgreSQL
+#     matrix and promotion list to src/config/jpa/release-registry.json.
+#   * fileserver-certification.yml — FileserverDocumentationCoverageTest requires every job id named
+#     in docs/fileserver/support-matrix.md to be defined in a `.github/workflows/fileserver-*.yml`.
+#     It is named "certification" rather than "release" on purpose: it certifies a storage topology
+#     and a support matrix, it deploys nothing, and the CI/CD boundary in docs/ci-cd/boundary.md
+#     says GitHub Actions does not deploy.
+# Folding either one in needs its src-side test (and, for fileserver, the support document) changed
+# in the same commit. Until then the image job below cannot wait on them — a stated gap.
+#
+# The image job DOES now wait on the evidence jobs in this file, which is new: while the image build
+# lived in its own workflow it could publish while a sibling suite was still running or already red,
+# because `needs:` does not reach across workflows.
+
+on:
+  push:
+    tags:
+      - "v*"
+
+permissions:
+  contents: read
+
+# Never cancel a release in flight. A half-pushed manifest is worse than a slow one, and two runs
+# for the same tag would race for the same registry tags.
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: false
+
+jobs:
+  # The architecture-wide verification, once. Each of the four release workflows this file replaces
+  # ran `verifyCleanArchitectureDependencies` on its own runner, and three of them also ran the
+  # bootstrap architecture suite; the answers were identical because the input was one commit.
+  architecture-and-surface:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Verify architecture boundaries and the published surfaces
+        working-directory: src
+        run: >-
+          ./gradlew
+          :verifyCleanArchitectureDependencies
+          :verifyPublicPathSnapshot
+          :app-bootstrap:verifyEnvKeys
+          :app-bootstrap:test --tests 'dev.caskeleton.bootstrap.architecture.*'
+
+          --stacktrace
+
+  # Every web lane that `check` cannot reach. webCrossStackParityTest depends on `test`,
+  # webJettyCompatTest and webFluxContractTest, so naming it runs all four — which is what
+  # web-advanced-release.yml spent a separate 90-minute job doing by naming the three by hand.
+  #
+  # webAdvancedTest is here rather than in a nightly of its own. Its tests run inside
+  # `:adapter:inbound:web:test` by design, so the lane adds exactly one thing: it fails closed when
+  # the `web-advanced` tag selects nothing. That is worth asserting at a release and is not worth a
+  # workflow file and a runner every night.
+  web-stable-release-gate:
+    runs-on: ubuntu-latest
+    timeout-minutes: 90
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run every web lane, Stable and Advanced
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:web:webCrossStackParityTest
+          :adapter:inbound:web:webNginxProxyTest
+          :adapter:inbound:web:webAdvancedTest
+
+          --stacktrace
+      - name: Publish the release evidence
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: web-release-evidence
+          path: |
+            src/adapter/inbound/web/build/web-contract-parity/
+            src/adapter/inbound/web/build/reports/tests/
+          if-no-files-found: error
+
+  websocket-stable-release-gate:
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run every websocket lane, Stable and Advanced
+        working-directory: src
+        run: >-
+          ./gradlew
+          :adapter:inbound:websocket:test
+          :adapter:inbound:websocket:websocketJettyTest
+          :adapter:inbound:websocket:websocketNginxTest
+          :adapter:inbound:websocket:websocketTransportQualificationTest
+          :adapter:inbound:websocket:websocketAdvancedTest
+
+          --stacktrace
+      - name: Publish the release evidence
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: websocket-release-evidence
+          path: src/adapter/inbound/websocket/build/reports/tests/
+          if-no-files-found: error
+
+  # The three gRPC certification lanes. Their tests already run on every pull request — the
+  # `grpc-inprocess`, `grpc-netty` and `grpc-fault` tags are NOT excluded from
+  # `:grpc:grpc-testkit:test` (only `grpc-performance` is), and that task runs inside the root
+  # `check`. So this job adds exactly what the web and WebSocket Advanced lanes above add: the lane
+  # fails closed when its tag selects nothing, which is the one thing a tag-filtered suite inside
+  # `test` cannot tell you. A renamed or deleted @Tag would otherwise leave the in-process,
+  # transport and fault evidence grades claiming coverage that stopped existing.
+  #
+  # Release rather than nightly, for the same reason web-stable-release-gate is: these lanes need no
+  # container and no fixed cadence — grpcNettyContractTest opens an ephemeral socket, not a broker —
+  # so the guard is worth asserting once per tag and is not worth a runner every night.
+  #
+  # grpcPerformanceTest is deliberately absent. The leaf excludes it from `test` and says why: "a
+  # measurement in the release gate is a flaky test on a shared CI runner; it runs when somebody asks
+  # for it, by name." It is recorded as a manual entrypoint in the repository README instead.
+  grpc-stable-release-gate:
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the gRPC Stable certification lanes
+        working-directory: src
+        run: >-
+          ./gradlew -p optional-platforms
+          :grpc:grpc-testkit:grpcInProcessContractTest
+          :grpc:grpc-testkit:grpcNettyContractTest
+          :grpc:grpc-testkit:grpcFaultTest
+          --stacktrace
+      - name: Publish the gRPC release evidence
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # actions/upload-artifact@v4.6.2
+        with:
+          name: grpc-release-evidence
+          path: src/grpc/grpc-testkit/build/reports/tests/
+          if-no-files-found: warn
+
+  # Each gate runs as its own single-line `./gradlew <task>` step so that a failure names the gate
+  # rather than a folded command. The architecture dependency gate that used to end this list is now
+  # architecture-and-surface above; it was the fourth copy of the same invocation.
+  httpclient-release-gate:
+    runs-on: ubuntu-latest
+    timeout-minutes: 60
+    defaults:
+      run:
+        working-directory: src
+    env:
+      # A project property rather than a command-line flag, so each run command stays a plain,
+      # verifiable task invocation while the machine-dependent bounds are still asserted.
+      GRADLE_OPTS: -Dorg.gradle.project.performance.assertions.enabled=true
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: ./.github/actions/setup-gradle-java
+      - name: Run the HTTP client release qualification graph
+        run: >-
+          ./gradlew
+          :adapter:outbound:httpclient:test
+          :adapter:outbound:httpclient:spring62ApiSurfaceScan
+          :adapter:outbound:httpclient:spring70CompatibilityTest
+          :adapter:outbound:httpclient:httpClientStableContractTest
+          :adapter:outbound:httpclient:httpClientSecurityTest
+          :adapter:outbound:httpclient:httpClientBlockHoundTest
+          :adapter:outbound:httpclient:httpClientFailureInjectionTest
+          :adapter:outbound:httpclient:httpClientPerformanceTest
+          --stacktrace
+
+
+  httpclient-documentation:
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # actions/setup-python@v5.6.0
+        with:
+          python-version: '3.12'
+      - name: Verify documentation matches the code
+        run: python3 scripts/verify-httpclient-docs.py
+
+  app-image-release:
+    needs:
+      - architecture-and-surface
+      - web-stable-release-gate
+      - websocket-stable-release-gate
+      - grpc-stable-release-gate
+      - httpclient-release-gate
+      - httpclient-documentation
+    # Job-level, because a job that declares `permissions:` replaces the workflow set entirely: this
+    # is the only job that writes anything anywhere, and `packages: write` stops at its boundary.
+    permissions:
+      contents: read
+      packages: write
+    runs-on: ubuntu-latest
+    timeout-minutes: 90
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # actions/checkout@v4.2.2
+      # The builder stage inside src/Dockerfile runs this repository's Gradle wrapper to produce the
+      # JAR that becomes the image. Validating the wrapper here checks the thing that is about to
+      # execute, before it executes, rather than after an image already exists.
+      # The tag is the release identity; everything below derives from it. A tag that does not parse
+      # stops the release here, rather than producing an image named after whatever ref happened to
+      # be checked out.
+      #
+      # GHCR rejects an uppercase path, and this repository's owner is mixed case — the naive
+      # `ghcr.io/${{ github.repository }}` fails at push time with a message about the manifest
+      # rather than about the case, so the lowercasing is explicit and the result is asserted.
+      - name: Resolve the release coordinates
+        env:
+          CONFIGURED_IMAGE_NAME: ${{ vars.APP_IMAGE_NAME }}
+        run: |
+          set -euo pipefail
+          readonly REGISTRY='ghcr.io'
+          if [[ "${GITHUB_REF_TYPE}" != 'tag' ]]; then
+            echo "::error::container-release runs only for a release tag; ref type was ${GITHUB_REF_TYPE}"
+            exit 1
+          fi
+          release_tag="${GITHUB_REF_NAME}"
+          # Bare MAJOR.MINOR.PATCH, because src/build.gradle's release-version guard refuses a
+          # pre-release or build suffix and the image tag must be the same string the JAR reports.
+          if [[ ! "${release_tag}" =~ ^v([0-9]+\.[0-9]+\.[0-9]+)$ ]]; then
+            echo "::error::release tag must be vMAJOR.MINOR.PATCH with no suffix; got '${release_tag}'"
+            exit 1
+          fi
+          release_version="${BASH_REMATCH[1]}"
+          owner_path="$(printf '%s' "${GITHUB_REPOSITORY_OWNER}" | tr '[:upper:]' '[:lower:]')"
+          image_name="${CONFIGURED_IMAGE_NAME:-${owner_path}/caskeleton}"
+          image_repository="${REGISTRY}/${image_name}"
+          if [[ "${image_repository}" != "${image_repository,,}" ]]; then
+            echo "::error::image repository must be lowercase; got '${image_repository}'"
+            exit 1
+          fi
+          if [[ "${image_repository}" =~ [[:space:]] || "${image_repository}" == *:* ]]; then
+            echo "::error::image repository must carry no tag and no whitespace; got '${image_repository}'"
+            exit 1
+          fi
+          {
+            printf 'REGISTRY=%s\n' "${REGISTRY}"
+            printf 'RELEASE_VERSION=%s\n' "${release_version}"
+            printf 'BUILD_VERSION=%s+%s\n' "${release_version}" "${GITHUB_SHA}"
+            printf 'IMAGE_REPOSITORY=%s\n' "${image_repository}"
+            printf 'IMAGE_VERSION_TAG=%s\n' "${release_version}"
+            printf 'IMAGE_REVISION_TAG=sha-%s\n' "${GITHUB_SHA}"
+            printf 'SOURCE_URL=%s/%s\n' "${GITHUB_SERVER_URL}" "${GITHUB_REPOSITORY}"
+          } >> "${GITHUB_ENV}"
+          printf 'container-release: %s -> %s:%s and %s:sha-%s\n' \
+            "${release_tag}" "${image_repository}" "${release_version}" \
+            "${image_repository}" "${GITHUB_SHA}"
+      # Byte-identical to the install in dependency-vulnerability.yml, deliberately: the same
+      # checksum-pinned binary at the same version scans the filesystem and the image, so the two
+      # gates cannot disagree because one of them silently moved to a newer database schema.
+      #
+      # This repository installs its scanner rather than calling a scanner action, which is why no
+      # third-party action appears in this workflow: a pinned tarball with an asserted SHA-256 is a
+      # supply-chain claim that can be checked offline, and an action pinned to a commit is not.
+      - name: Install pinned Trivy under RUNNER_TEMP
+        env:
+          TRIVY_DOWNLOAD_BASE_URL: ${{ vars.TRIVY_DOWNLOAD_BASE_URL }}
+        run: |
+          set -euo pipefail
+          readonly TRIVY_VERSION='0.71.2'
+          readonly TRIVY_SHA256_AMD64='0510e71e2fd39bf863856d499c8dc19feb4e7336546394c502a8f5cc7ab27460'
+          readonly TRIVY_SHA256_ARM64='fe1c7106e15a5365d485b098a8c338f91e3b7ba71cb0e4963b98a3a098763cfc'
+          readonly DOWNLOAD_BASE_URL="${TRIVY_DOWNLOAD_BASE_URL:-https://github.com/aquasecurity/trivy/releases/download/v${TRIVY_VERSION}}"
+          case "${RUNNER_ARCH:-X64}" in
+            X64)
+              asset_arch='64bit'
+              expected_sha256="${TRIVY_SHA256_AMD64}"
+              ;;
+            ARM64)
+              asset_arch='ARM64'
+              expected_sha256="${TRIVY_SHA256_ARM64}"
+              ;;
+            *)
+              echo "::error::unsupported runner architecture: ${RUNNER_ARCH:-unknown}"
+              exit 1
+              ;;
+          esac
+          install_dir="${RUNNER_TEMP}/trivy-${TRIVY_VERSION}"
+          archive="${RUNNER_TEMP}/trivy-${TRIVY_VERSION}.tar.gz"
+          mkdir -p "${install_dir}"
+          curl --fail --show-error --silent --location --retry 3 \
+            --proto '=https' --tlsv1.2 \
+            "${DOWNLOAD_BASE_URL}/trivy_${TRIVY_VERSION}_Linux-${asset_arch}.tar.gz" \
+            --output "${archive}"
+          printf '%s  %s\n' "${expected_sha256}" "${archive}" | sha256sum -c -
+          tar -xzf "${archive}" -C "${install_dir}" trivy
+          chmod 0755 "${install_dir}/trivy"
+          printf '%s\n' "${install_dir}" >> "${GITHUB_PATH}"
+      # SOURCE_DATE_EPOCH is the commit time, not the wall clock, so the image metadata is a function
+      # of the commit rather than of when the runner happened to pick the job up. Verified locally,
+      # and worth stating exactly because it is easy to overclaim: BuildKit uses it for the image
+      # config `created` field and for every history timestamp — both came back as the commit time —
+      # and it does NOT rewrite file mtimes inside the layers. Those still carry the build time, so
+      # two builds of the same commit agree on metadata but their layer digests still differ.
+      # Byte-identical layers additionally need `--output type=image,rewrite-timestamp=true`, which
+      # needs the containerd image store; that is a runner-capability change, not a flag to add
+      # untested to the one job that publishes releases.
+      #
+      # The OCI `created` label comes from the same commit for the same reason: `date -u` there would
+      # have made every rebuild a different image for no reason anybody could see.
+      #
+      # Both base images are already digest-pinned inside src/Dockerfile, and so is the Dockerfile
+      # frontend in its `# syntax` directive, so nothing in this build resolves a floating tag.
+      - name: Build the release image
+        run: |
+          set -euo pipefail
+          SOURCE_DATE_EPOCH="$(git log -1 --format=%ct)"
+          export SOURCE_DATE_EPOCH
+          created="$(git log -1 --format=%cI)"
+          printf 'SOURCE_DATE_EPOCH=%s (%s)\n' "${SOURCE_DATE_EPOCH}" "${created}"
+          DOCKER_BUILDKIT=1 docker build \
+            --file src/Dockerfile \
+            --tag "${IMAGE_REPOSITORY}:${IMAGE_VERSION_TAG}" \
+            --tag "${IMAGE_REPOSITORY}:${IMAGE_REVISION_TAG}" \
+            --build-arg RELEASE_VERSION="${RELEASE_VERSION}" \
+            --build-arg BUILD_VERSION="${BUILD_VERSION}" \
+            --build-arg GIT_SHA="${GITHUB_SHA}" \
+            --build-arg SOURCE_URL="${SOURCE_URL}" \
+            --label org.opencontainers.image.created="${created}" \
+            src
+          docker image inspect \
+            --format 'built {{.Id}} ({{.Size}} bytes, {{len .RootFS.Layers}} layers)' \
+            "${IMAGE_REPOSITORY}:${IMAGE_VERSION_TAG}"
+      # Generated before the blocking scan, and uploaded before it too, so the inventory of what is
+      # in the image survives the run that refuses to publish it. An SBOM you only get on a green
+      # build is an SBOM you cannot use to answer "what was in the one that failed".
+      - name: Generate the image SBOM
+        run: |
+          set -euo pipefail
+          trivy image \
+            --format cyclonedx \
+            --scanners license \
+            --output image-sbom.cdx.json \
+            "${IMAGE_REPOSITORY}:${IMAGE_VERSION_TAG}"
+          test -s image-sbom.cdx.json
+      - name: Upload the image SBOM
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # actions/upload-artifact@v7.0.1
+        with:
+          name: container-release-sbom
+          path: image-sbom.cdx.json
+          if-no-files-found: error
+          retention-days: 90
+      # The same policy dependency-vulnerability.yml applies to the filesystem, applied to the thing
+      # that actually ships: CRITICAL and HIGH block, everything else is reported. The filesystem
+      # scan cannot see the base image's OS packages, which is most of an image's attack surface, so
+      # a green trivy-fs has never been evidence about the artifact.
+      #
+      # --ignorefile is mandatory here as everywhere: .trivyignore.yaml is the single suppression
+      # source. Each entry carries a rationale and an expiry by policy, reviewed through CODEOWNERS
+      # (.github/dependency-vulnerability-policy.md); an inline --skip or a second ignore file would
+      # be a suppression nobody reviews.
+      #
+      # This step is the reason `docker push` is further down. A vulnerable image that was pushed and
+      # then reported is already pullable by everything that watches the tag.
+      - name: Block High and Critical vulnerabilities in the release image
+        run: |
+          set -euo pipefail
+          trivy image \
+            --scanners vuln,license \
+            --severity CRITICAL,HIGH \
+            --exit-code 1 \
+            --ignorefile .trivyignore.yaml \
+            "${IMAGE_REPOSITORY}:${IMAGE_VERSION_TAG}"
+      - name: Report Medium and Low vulnerabilities in the release image
+        run: |
+          set -euo pipefail
+          trivy image \
+            --scanners vuln,license \
+            --severity MEDIUM,LOW \
+            --exit-code 0 \
+            --ignorefile .trivyignore.yaml \
+            "${IMAGE_REPOSITORY}:${IMAGE_VERSION_TAG}"
+      - name: Sign in to the container registry
+        env:
+          REGISTRY_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          set -euo pipefail
+          printf '%s' "${REGISTRY_TOKEN}" \
+            | docker login "${REGISTRY}" --username "${GITHUB_ACTOR}" --password-stdin
+      # Two tags, one digest. The semver tag is what a human reads and what a release note cites; the
+      # sha- tag is the one that can never be moved to different content, because the git SHA it
+      # names is the only commit that can produce it.
+      #
+      # Neither is what a manifest should pin. Both are mutable names in a registry: a later push can
+      # point `1.2.3` at something else, and nothing about a tag tells a cluster it did not. The
+      # digest recorded below is immutable by construction, and it is the field the GitOps repository
+      # pins — the tags exist so a person can find the digest, not so a cluster can resolve one.
+      - name: Push the release and revision tags
+        run: |
+          set -euo pipefail
+          docker push "${IMAGE_REPOSITORY}:${IMAGE_VERSION_TAG}"
+          docker push "${IMAGE_REPOSITORY}:${IMAGE_REVISION_TAG}"
+          # awk rather than `grep | head`, deliberately. Under `set -e` with `pipefail`, a grep that
+          # matches nothing exits 1 and kills the step right here — so the explicit check below,
+          # and its message, would never run and the failure would surface as a bare exit code.
+          # awk exits 0 whether or not it matched, which leaves the empty case for us to report.
+          pinned_reference="$(
+            docker image inspect \
+              --format '{{range .RepoDigests}}{{println .}}{{end}}' \
+              "${IMAGE_REPOSITORY}:${IMAGE_VERSION_TAG}" \
+              | awk -v prefix="${IMAGE_REPOSITORY}@sha256:" \
+                  'index($0, prefix) == 1 { print; exit }'
+          )"
+          if [[ -z "${pinned_reference}" ]]; then
+            echo "::error::no registry digest for ${IMAGE_REPOSITORY} after push"
+            exit 1
+          fi
+          printf 'PINNED_REFERENCE=%s\n' "${pinned_reference}" >> "${GITHUB_ENV}"
+          printf 'container-release: pushed %s\n' "${pinned_reference}"
+      # The handoff to the GitOps repository, in a form a person and a script can both read. It is
+      # written to the job summary as well as to an artifact because the summary is where somebody
+      # looks first and the artifact is what survives the ninety days a release audit asks about.
+      - name: Record the immutable image reference
+        run: |
+          set -euo pipefail
+          digest="${PINNED_REFERENCE#*@}"
+          {
+            printf 'release_tag: %s\n' "${GITHUB_REF_NAME}"
+            printf 'git_sha: %s\n' "${GITHUB_SHA}"
+            printf 'image_repository: %s\n' "${IMAGE_REPOSITORY}"
+            printf 'version_tag: %s\n' "${IMAGE_VERSION_TAG}"
+            printf 'revision_tag: %s\n' "${IMAGE_REVISION_TAG}"
+            printf 'digest: %s\n' "${digest}"
+            printf 'pinned_reference: %s\n' "${PINNED_REFERENCE}"
+          } > image-release.txt
+          {
+            printf '### container-release\n\n'
+            printf 'Pin this in the GitOps manifest as the container image:\n\n'
+            printf '```\n%s\n```\n\n' "${PINNED_REFERENCE}"
+            printf -- '- release tag: `%s`\n' "${GITHUB_REF_NAME}"
+            printf -- '- version tag: `%s:%s`\n' "${IMAGE_REPOSITORY}" "${IMAGE_VERSION_TAG}"
+            printf -- '- revision tag: `%s:%s`\n' "${IMAGE_REPOSITORY}" "${IMAGE_REVISION_TAG}"
+          } >> "${GITHUB_STEP_SUMMARY}"
+          cat image-release.txt
+      - name: Upload the immutable image reference
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # actions/upload-artifact@v7.0.1
+        with:
+          name: container-release-image-reference
+          path: image-release.txt
+          if-no-files-found: error
+          retention-days: 90
@@ -1,2 +1,12 @@
 .vscode/
 src/**/bin/
+.claude/
+
+# Operator input, not a build input. The examples beside it are the tracked contract;
+# verifyEnvKeys reads the registry, the profile YAMLs and .env.example, never a real one.
+src/.env*
+!src/.env.example
+!src/.env.local.example
+
+# Written per run by the runtime-smoke wrapper; never committed.
+src/.env.lane
@@ -1,13 +1,21 @@
 # Structured Trivy suppression baseline.
 #
-# This repository-root file is the only CI suppression source. Every future entry must include:
-#   id: advisory, license, misconfiguration, or secret identifier
-#   statement: non-empty accepted-risk or false-positive rationale
-#   expired_at: future YYYY-MM-DD no more than 90 days from review
+# This repository-root file is the only CI suppression source. Every Trivy invocation must name it
+# with `--ignorefile .trivyignore.yaml`; ad-hoc ignore files and inline bypasses are not allowed.
 #
-# `verifyTrivyignore` enforces those fields and the expiry window. CODEOWNERS supplies the separate
-# reviewer control. Every Trivy invocation must also name this file with
-# `--ignorefile .trivyignore.yaml`; do not add ad-hoc ignore files or inline bypasses.
+# Every entry must carry:
+#   id:         advisory, license, misconfiguration, or secret identifier
+#   statement:  non-empty accepted-risk or false-positive rationale
+#   expired_at: future YYYY-MM-DD, no more than 90 days from review
+#
+# Enforced by review, not by a build task. `verifyTrivyignore` used to be a 105-line hand-written
+# YAML parser in the root build — indentation tracking, inline-scalar handling, quote stripping — and
+# what it guarded was this file, which has been empty since it was created. A suppression is added by
+# a human and merged by a CODEOWNERS reviewer (.github/dependency-vulnerability-policy.md); that
+# reviewer is the control, and a parser that has never seen an entry is not a second one.
+#
+# If this file ever carries entries and they start drifting, that is the moment to automate the
+# check — against real entries, with a real YAML library. Not before.

 vulnerabilities: []
 licenses: []
@@ -49,11 +49,22 @@ root `CLAUDE.md`는 이 목록의 동기화된 요약이다. 두 문서가 어

 ## Gradle 정책 권위

- `src/config/architecture/modules.json`: 정확히 19개 leaf의 ID, repository-relative 소스 경로,
-  Gradle path, 허용 production project dependency edge
- `src/settings.gradle`: registry를 fail-closed로 검증하고 등록된 Gradle project를 include/mapping
- `src/build.gradle`: 같은 registry를 읽는 `verifyCleanArchitectureDependencies`와 그 밖의
-  architecture-wide verification task
+- `src/config/architecture/modules.json`: 등록된 모든 leaf의 ID, repository-relative 소스 경로,
+  Gradle path, 허용 production project dependency edge, 두 composition root의 실제 runtime
+  membership. leaf 목록과 그 개수의 SSOT는 registry다. 문서는 개수를 복제하지 않는다 —
+  산문에 적힌 숫자는 leaf가 추가되는 순간 drift하기 때문이다. 이제 이걸 강제하는 태스크는 없다:
+  `verifyDocumentedLeafCount`는 삭제됐다. 문서에 적힌 수가 틀린 것은 결함이지만 빌드를 실패시킬
+  사유는 아니고, 그 태스크는 모든 `CLAUDE.md`와 `build.gradle`을 정규식으로 훑는 파서였다.
+- `src/settings.gradle`: 16줄. `ca.architecture-registry` 설정 플러그인이 registry를 읽어
+  project를 include/mapping 한다. registry가 project 목록이 될 수 없는 경우(중복 ID, 저장소 밖
+  경로, 없는 디렉터리)만 여기서 실패한다. 허용되지 않는 edge 같은 아키텍처 규칙은
+  `verifyCleanArchitectureDependencies`가 답한다 — settings에서 죽으면 실행할 수 있는 태스크가
+  하나도 없다.
+- `src/build-logic/`: convention plugin. leaf는 `ca.java-library` / `ca.spring-library` /
+  `ca.platform-module` 중 자기 성격을 선언하고, 그 플러그인이 toolchain·락·정적분석·테스트
+  기본값을 준다. `ca.architecture`가 아키텍처 검증 태스크를 소유한다.
+- `src/build.gradle`: 루트 라이프사이클(`ci`, `releaseCheck`, `qualityCheck`,
+  `configContractCheck`, `integrationCheck`)과 버전/리비전

 작업 파일의 소유 leaf는 registry의 `source_path`로 판단하고 가장 가까운 `src/**/CLAUDE.md`를
 함께 읽는다. focused test는 registry의 `gradle_path`에서
@@ -101,7 +112,7 @@ commit 정책은 모든 플랫폼에서 `human-only`이며 agent는 stage/commit

 ## 모듈 책임

-19개 leaf 모듈의 ID, 실제 소스 경로, Gradle path, 허용 production 의존성은
+모든 leaf 모듈의 ID, 실제 소스 경로, Gradle path, 허용 production 의존성, runtime membership은
 `src/config/architecture/modules.json`이 SSOT다. focused test는 소유 leaf의 `gradle_path`에서
 파생한다. 이 문서는 leaf 목록을 복제하지 않고 family 책임만 정의한다. 작업 파일에서는 가장 가까운
 `src/**/CLAUDE.md`를 함께 읽는다.
@@ -172,16 +183,19 @@ Gradle 의존성 검증도 같은 registry를 읽는다. root 문서나 기억

 ```bash
 cd src
-./gradlew <owner-gradle-path>:test --console=plain
-./gradlew test
-./gradlew check    # check 가 verifyCleanArchitectureDependencies + verifyEnvKeys 2종을 전이 실행한다 (src/build.gradle)
-./gradlew verifyCleanArchitectureDependencies
+./gradlew <owner-gradle-path>:check --console=plain   # 그 leaf만: 컴파일·테스트·포맷·스타일·ErrorProne
+./gradlew check                                      # 모든 leaf의 check
+./gradlew architectureCheck                          # 의존 방향·런타임 멤버십·application-core 순수성
+./gradlew qualityCheck                               # SpotBugs + FindSecBugs (leaf check에는 없다)
+./gradlew ci                                         # PR 게이트 = 위 셋 + configContractCheck
 ./gradlew verifyPublicPathSnapshot
-./gradlew verifyEnvKeys
+./gradlew :app-bootstrap:verifyEnvKeys
 ```

+leaf의 `check`는 그 leaf만 검사한다. 저장소 전체 질문은 이름이 따로 있는 루트 태스크가 답한다.
+
 소유 leaf의 정확한 Gradle path는 `src/config/architecture/modules.json`에서 읽고 focused test
-명령을 파생한다. root 문서에 19개 명령 목록을 복제하지 않는다.
+명령을 파생한다. root 문서에 leaf별 명령 목록을 복제하지 않는다.

 ## 설정과 런타임

@@ -1,6 +1,6 @@
 # CLAUDE.md

-Repository guidance for the Java 21 + Spring Boot 4.0.0 Clean Architecture template.
+Repository guidance for the Java 21 + Spring Boot 4.0.8 Clean Architecture template.

 ## Prime Directive

@@ -21,8 +21,10 @@ If this summary drifts from `AGENTS.md`, `AGENTS.md` wins and this summary must

 ## Gradle policy authorities

- `src/config/architecture/modules.json`: exactly 19 leaf identities, repository-relative source
-  paths, Gradle paths, and allowed production project dependency edges.
+- `src/config/architecture/modules.json`: every registered leaf identity, repository-relative source
+  paths, Gradle paths, allowed production project dependency edges, and the exact runtime
+  memberships of both composition roots. The registry owns the leaf list and its size; no document
+  restates the count, because a number written in prose drifts the moment a leaf is added.
 - `src/settings.gradle`: fail-closed registry validation, project inclusion, and directory mapping.
 - `src/build.gradle`: `verifyCleanArchitectureDependencies` and the other architecture-wide
  verification tasks.
@@ -42,8 +44,10 @@ count.

 ## Module families

-`src/config/architecture/modules.json` owns the complete 19-leaf list. Root guidance summarizes
-families; the nearest `src/**/CLAUDE.md` owns local rules.
+`src/config/architecture/modules.json` owns the complete leaf list. Root guidance summarizes
+families; the nearest `src/**/CLAUDE.md` owns local rules. No task enforces this any more:
+`verifyDocumentedLeafCount` was deleted along with the other documentation-drift parsers. A stated
+count that disagrees with the registry is a defect, not a build failure — so do not state one.

 | Family | Responsibility | Stable dependency direction |
 | --- | --- | --- |
@@ -53,11 +57,22 @@ families; the nearest `src/**/CLAUDE.md` owns local rules.
 | `adapter:outbound:persistence-*` | JPA/PostgreSQL and MongoDB persistence adapters | application/domain/shared contracts as registered |
 | `adapter:outbound:*` | support, messaging, cache, notification, storage, file, HTTP client, identifier capabilities | application/domain/shared and registered support edge |
 | `shared-contract` | Skeleton-wide operational contracts | Java stdlib only |
+| `messaging:*` | Vendored messaging platform: a product with its own API, SPI, adapters and composition boundary, not a layer of this application | `messaging:*` only — it depends on no `domain-core`, `application-core`, or `shared-contract` type |
 | `sample-portfolio` | Fixture/reference consumer | registered runtime leaves; never a production dependency |
 | `app-bootstrap` | Spring Boot entrypoint and composition root | registered runtime leaves |

+The `messaging:*` family is the one entry that is not a Clean Architecture layer, and it is listed so
+that the exception is stated rather than inferred from a directory. It is a vendored library — its
+own `*-api` leaves are its ports, its broker leaves are its adapters, its starter is its composition
+root — and the messaging module review (`docs/reviews/2026-08-14-messaging-module-code-review.md`
+MSG-023 §6.2) chose that layout deliberately over folding it into `adapter:outbound:*`. This
+application is supposed to reach it the way it reaches any library: through an application-owned port
+satisfied by an anti-corruption bridge in `adapter:outbound:messaging`. That bridge does not exist
+yet (MSG-015), so today the composition root wires the starter directly; `src/messaging/CLAUDE.md`
+holds the detail.
+
 Never infer an individual leaf's Gradle path, allowed dependency, or test command from this table.
-Read its `gradle_path` and `allowed_dependencies` from
+Read its `gradle_path`, `allowed_dependencies`, and `runtime_memberships` from
 `src/config/architecture/modules.json`; derive the focused test from that Gradle path.

 ## Layer workflow
@@ -91,14 +106,21 @@ From `src/`, read the owning leaf's `gradle_path` from
 Architecture-wide commands:

 ```bash
-./gradlew verifyCleanArchitectureDependencies --console=plain
-./gradlew :app-bootstrap:test --tests '*CleanArchitectureTest' --console=plain
+./gradlew architectureCheck --console=plain
+./gradlew :app-bootstrap:test --tests 'dev.caskeleton.bootstrap.architecture.*' --console=plain
 ./gradlew verifyPublicPathSnapshot --console=plain
-./gradlew verifyEnvKeys --console=plain
+./gradlew :app-bootstrap:verifyEnvKeys --console=plain
 ```

-Use public-path and env-key checks only when their surfaces changed. Full `test` or `check` requires
-the controller's workflow authorization.
+A leaf's `check` covers that leaf only — compile, its tests, Spotless, Checkstyle, Error Prone.
+Repository-wide questions have their own names: `architectureCheck` (dependency direction, runtime
+membership, application-core purity, Git-carryable sources), `qualityCheck` (SpotBugs, FindSecBugs),
+`configContractCheck` (the environment contract), `integrationCheck` (the declared strict test
+lanes). `ci` is check + architectureCheck + qualityCheck + configContractCheck; `releaseCheck` adds
+provenance, archive hygiene and the public-path snapshot.
+
+Use public-path and env-key checks only when their surfaces changed. Full `test`, `check` or `ci`
+requires the controller's workflow authorization.

 ## Advisory and reporting

@@ -56,6 +56,23 @@ docker compose -f docker-compose.yml -f docker-compose.local.yml down

 `src/.env`는 커밋된 안전 기본값이라 별도 `.env.example`을 만들지 않습니다. 전체 환경 변수 목록과 조정 시점은 [src/README.md](src/README.md)와 [docs/registries/env-keys.yaml](docs/registries/env-keys.yaml)에 있습니다.

+### 프로파일별 데이터스토어
+
+`bootstrap`은 컨테이너 경로(PostgreSQL)를 검증하는 첫 실행 진입점입니다. 일상 개발은 Docker 없이 돌리는 `local` 프로파일이며, 이때 데이터스토어는 H2 in-memory입니다.
+
+```bash
+cd src
+./gradlew :app-bootstrap:bootRun
+```
+
+| 프로파일 | 데이터스토어 | 스키마 소유자 |
+| --- | --- | --- |
+| `local` (bootRun 기본) | H2 in-memory | Hibernate `create-drop` |
+| `dev` | PostgreSQL | Flyway |
+| `prod` | PostgreSQL | Flyway |
+
+`local`은 wiring과 애플리케이션 동작을 검증하고, migration과 vendor 동작은 검증하지 않습니다. 프로파일별 설정은 [src/app-bootstrap/src/main/resources/](src/app-bootstrap/src/main/resources/)의 `application-{local,dev,prod}.yml`이, 상세 설명은 [src/README.md](src/README.md)가 소유합니다.
+
 ## 새 프로젝트로 시작하기

 이 저장소를 새 서비스의 출발점으로 쓸 때 핵심 단계는 다음과 같습니다. 전체 체크리스트는 [AGENTS.md](AGENTS.md)의 "템플릿 재사용 체크리스트"에 있습니다.
@@ -73,14 +90,20 @@ docker compose -f docker-compose.yml -f docker-compose.local.yml down
 3. `CaSkeletonApplication`을 새 애플리케이션 이름으로 바꾸고, 목표 도메인의 엔티티·repository port·유스케이스·adapter를 production 모듈에 추가합니다. 예시 코드는 `sample-portfolio`에만 둡니다.
 4. 모듈 이름과 경계는 그대로 유지합니다.

-검증은 sample-on과 sample-off를 모두 통과시킵니다.
+검증은 먼저 composition root의 빠른 테스트와 sample-off 계약을 확인합니다. 루트에서
+`./gradlew test`를 호출하면 등록된 모든 하위 프로젝트의 `test`를 실행하므로 일상적인 로컬
+피드백 명령으로 사용하지 않습니다. 저장소 전체 qualification은 CI 또는 명시적인 `ci` task가
+담당합니다.

 ```bash
 cd src
-./gradlew test
+./gradlew :app-bootstrap:test
 ./gradlew :app-bootstrap:sampleOffTest
+./gradlew architectureCheck
 ```

+병합 전 저장소 전체 검증이 필요하면 `./gradlew ci`를 실행합니다.
+
 `sample-portfolio`는 템플릿이 유지하는 fixture/reference 모듈이라 production 모듈이 의존하지 않고, runtime에 sample bean이나 endpoint를 넣지 않습니다. 다운스트림 fork에서 fixture가 더 필요 없을 때만 sample-off 테스트를 통과시킨 뒤 정리합니다.

 ## 아키텍처 규칙과 검증
@@ -105,6 +128,35 @@ cd src

 두 검증 축은 [ci-quality-gates.yml](.github/workflows/ci-quality-gates.yml)의 release gate에 연결되어, 규칙 위반이 병합·릴리스를 막습니다.

+## 수동 전용 Gradle 태스크
+
+아래 세 태스크는 **어떤 워크플로도 실행하지 않으며, 그게 의도다.** 자동 실행이 틀린 이유를 각각
+적어 둔다.
+
+여기 적힌 태스크 이름이 실재하는지 검사하던 `verifyReadmeCommands`는 삭제했다. 그건 이 문서의
+```bash 블록을 직접 파싱해 `./gradlew`·`docker compose`·`make` 토큰을 실제 태스크 그래프와 대조하는
+Markdown 명령 파서였고, 그 결과 "README에 무엇을 쓸 수 있는가"가 그 파서가 읽을 수 있는 문법의
+함수가 됐다. 문서와 코드가 어긋나는 것은 결함이지만, 빌드를 실패시켜서 고칠 일은 아니다.
+
+```bash
+cd src
+./gradlew :grpc:grpc-testkit:grpcPerformanceTest
+./gradlew :sample-portfolio:openapiCheckSnapshot -PapproveOpenApiChange
+./gradlew :app-bootstrap:sampleOffCompile
+```
+
+- `grpcPerformanceTest` — latency percentile·saturation·drain budget을 **측정**한다. 공유 CI
+  runner의 측정값은 흔들리고, 흔들리는 게이트는 결국 꺼진다. leaf `build.gradle`이 이 태스크의
+  태그를 `test`에서 제외하는 이유도 같다. 성능 회귀가 의심될 때 사람이 이름으로 부른다.
+- `openapiCheckSnapshot` — 드리프트 검사 자체는 이미 자동으로 돈다. 이 태스크가 감싸는
+  `OpenApiDriftContractTest`는 `:sample-portfolio:test`의 일부이고, 그건 `check` 안이며 stage 1에서
+  실행된다. 이 태스크의 고유한 역할은 `-PapproveOpenApiChange`로 **커밋된 스냅샷을 다시 만드는 것**
+  — 의도된 API 변경을 사람이 승인하는 지점이다. 자동으로 돌리면 승인이 승인이 아니게 된다.
+- `sampleOffCompile` — `sampleOffTest` 소스셋을 **컴파일만** 한다. CI가 돌리는
+  `:app-bootstrap:sampleOffTest`(stage 1, `ci-quality-gates.yml`의 `sample-off` 잡)는 같은 소스셋을
+  컴파일한 뒤 실행까지 하므로, CI에 따로 넣으면 진부분집합을 한 번 더 도는 것이다. 남겨 둔 이유는
+  sample 제거 작업 중 테스트를 기다리지 않고 컴파일만 빠르게 확인하는 로컬 루프가 실재하기 때문이다.
+
 ## 더 알아보기

 - 빌드·검증 게이트·환경 변수 상세: [src/README.md](src/README.md)
@@ -18,11 +18,30 @@ services:
  app:
    # Relax read-only constraint for local development.
    read_only: false
-    tmpfs: []       # no tmpfs in dev; rely on normal writable rootfs
+    # `!override`, not a plain empty list. An empty sequence merges with the base sequence rather
+    # than replacing it, so the base's /var/tmp/heap tmpfs survived and collided with the bind mount
+    # below — Compose refuses to have the same target twice and will not silently pick one. That is
+    # the right refusal: a heap dump written into a tmpfs dies with the container that produced it,
+    # which is the one moment somebody wants the file.
+    #
+    # `!override` needs Compose >= 2.24.4. Whether the collision is actually gone is checked in the
+    # merged model rather than assumed from this line.
+    tmpfs: !override []
    # More memory for dev profiling / heap dumps.
    mem_limit: 1g
    memswap_limit: 1g
    environment:
+      # Explicit, not inherited. A Compose profile selects services; it says nothing about which
+      # environment the application believes it is in, and the two drifting is how a dev stack ends
+      # up running local's settings.
+      SPRING_PROFILES_ACTIVE: "dev"
+      # The datasource address, owned here like the local and prod-smoke overlays own theirs. It was
+      # the only one of the three missing, and the gap was invisible while the qualification wrapper
+      # supplied a URL to every lane: the dev stack ran on a value that came from the test harness
+      # rather than from the file that describes the dev environment. With the wrapper no longer
+      # setting it — it was overriding prod's sslmode=verify-full URL — dev had none at all and
+      # Flyway was handed the literal string "${APP_DATASOURCE_URL}".
+      APP_DATASOURCE_URL: "jdbc:postgresql://db:5432/${POSTGRES_DB:-ca_skeleton}"
      TZ: "UTC"
      LANG: "C.UTF-8"
      LC_ALL: "C.UTF-8"
@@ -53,9 +72,30 @@ services:
    # Do not restart automatically so crash loops stay visible.
    restart: "no"
    # Optional: mount heap dump directory to host for dev analysis.
+    # The base declares /var/tmp/heap as a tmpfs, which is right for an ephemeral runtime and wrong
+    # for dev: a heap dump written into a tmpfs dies with the container that produced it, which is
+    # the one moment somebody wants the file. Compose refuses to have both, and correctly — it will
+    # not silently pick one — so the tmpfs list is replaced rather than appended to.
+    #
+    # `!override` needs Compose >= 2.24.4. An empty sequence is not assumed to delete the base
+    # sequence by itself; scripts/verify-compose-profile-contracts.sh checks mount-target uniqueness
+    # in the merged model, which is what actually proves the collision is gone.
    volumes:
      - type: bind
        source: ./tmp/heap-dumps
        target: /var/tmp/heap
        bind:
          create_host_path: true
+    # The same network the shared infrastructure lives on. The local overlay joins it and the dev
+    # overlay did not, so a dev lane that started PostgreSQL beside the application put the two on
+    # different networks: `UnknownHostException: db`, from a container that was running and healthy
+    # a metre away. Compose puts a service with no `networks:` on `default`, which is a network of
+    # its own making — so the omission reads as a working stack until something has to resolve a
+    # name across it.
+    networks:
+      - caskeleton-infra
+
+networks:
+  # Defined in docker-compose.infra.yml, where the services that share it live.
+  caskeleton-infra:
+    external: false
@@ -0,0 +1,503 @@
+# =============================================================================
+# Shared infrastructure, owned here and nowhere else.
+#
+# Environment overlays (local, dev, prod-smoke) describe how the application runs. This file
+# describes what it runs against. Keeping the two apart is why `local` could stop meaning "the app
+# plus a database" and start meaning "the app, with whichever services the lane asked for".
+#
+# Every service carries a Compose profile, so nothing here starts unless a lane names it. A profile
+# selects services; it never implies a Spring profile. The lane definitions live in
+# src/config/runtime/compose-profile-contracts.json, and scripts/verify-compose-profile-contracts.sh
+# checks this file against them.
+# =============================================================================
+
+services:
+  # ---- PostgreSQL --------------------------------------------------------------
+  db:
+    profiles:
+      - local-jpa
+      - local-messaging-outbox
+      - local-notification-ingest
+      - local-notification-serving
+      - local-notification-handoff
+      - shared-infra
+      - prod-smoke
+      - all-adapters
+    image: postgres:16-alpine
+    environment:
+      POSTGRES_DB: "${POSTGRES_DB:-ca_skeleton}"
+      POSTGRES_USER: "${APP_DATASOURCE_USERNAME:-ca_skeleton}"
+      POSTGRES_PASSWORD: "${APP_DATASOURCE_PASSWORD:-ca_skeleton}"
+      TZ: "UTC"
+    volumes:
+      - type: volume
+        source: caskeleton-db-data
+        target: /var/lib/postgresql/data
+    ports:
+      - "127.0.0.1:5433:5432"
+    networks:
+      - caskeleton-infra
+    healthcheck:
+      test:
+        [
+          "CMD-SHELL",
+          "pg_isready -U ${APP_DATASOURCE_USERNAME:-ca_skeleton} -d ${POSTGRES_DB:-ca_skeleton}",
+        ]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+      start_period: 30s
+
+  # ---- MongoDB -----------------------------------------------------------------
+  # A replica set of one. Single-node is still a replica set: transactions and change streams need
+  # one, and a standalone mongod that "works for reads" is a deployment that discovers the
+  # difference at the first transaction.
+  mongo:
+    profiles:
+      - local-mongo
+      - all-adapters
+    image: mongo:7
+    command: ["mongod", "--replSet", "rs0", "--bind_ip_all"]
+    volumes:
+      - type: volume
+        source: caskeleton-mongo-data
+        target: /data/db
+    networks:
+      - caskeleton-infra
+    healthcheck:
+      test: ["CMD", "mongosh", "--quiet", "--eval", "db.adminCommand('ping').ok"]
+      interval: 10s
+      timeout: 5s
+      retries: 10
+      start_period: 20s
+
+  mongo-rs-init:
+    profiles:
+      - local-mongo
+      - all-adapters
+    image: mongo:7
+    depends_on:
+      mongo:
+        condition: service_healthy
+    # Idempotent: rs.initiate() on an already-initiated set returns an error this swallows, so the
+    # lane can be re-run against a surviving volume without a manual reset.
+    command:
+      - mongosh
+      - --host
+      - mongo
+      - --quiet
+      - --eval
+      - >-
+        try { rs.initiate({_id:'rs0',members:[{_id:0,host:'mongo:27017'}]}) }
+        catch (e) { if (!/already initialized/i.test(e.message)) { throw e } }
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+  # ---- Kafka -------------------------------------------------------------------
+  kafka:
+    profiles:
+      - local-messaging
+      - local-messaging-outbox
+      - all-adapters
+    image: apache/kafka:3.8.0
+    environment:
+      KAFKA_NODE_ID: "1"
+      KAFKA_PROCESS_ROLES: "broker,controller"
+      KAFKA_LISTENERS: "PLAINTEXT://:9092,CONTROLLER://:9093"
+      KAFKA_ADVERTISED_LISTENERS: "PLAINTEXT://kafka:9092"
+      KAFKA_CONTROLLER_QUORUM_VOTERS: "1@kafka:9093"
+      KAFKA_CONTROLLER_LISTENER_NAMES: "CONTROLLER"
+      KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: "CONTROLLER:PLAINTEXT,PLAINTEXT:PLAINTEXT"
+      KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: "1"
+      KAFKA_TRANSACTION_STATE_LOG_REPLICATION_FACTOR: "1"
+      KAFKA_TRANSACTION_STATE_LOG_MIN_ISR: "1"
+    networks:
+      - caskeleton-infra
+    healthcheck:
+      test:
+        ["CMD-SHELL", "/opt/kafka/bin/kafka-broker-api-versions.sh --bootstrap-server kafka:9092"]
+      interval: 10s
+      timeout: 10s
+      retries: 12
+      start_period: 30s
+
+  # ---- Mailpit — the reference SMTP provider for notification serving ----------
+  mailpit:
+    profiles:
+      - local-notification-serving
+      - local-notification-handoff
+      - all-adapters
+    image: axllent/mailpit:v1.21
+    environment:
+      MP_SMTP_AUTH_ACCEPT_ANY: "1"
+      # MP_SMTP_AUTH_ALLOW_INSECURE is deliberately absent, and Mailpit refuses to start with both:
+      # "TLS cannot be required with --smtp-auth-allow-insecure". It existed to permit credentials
+      # over a plaintext connection, which is exactly what requiring STARTTLS removes the need for —
+      # any AUTH now happens inside the TLS session.
+      # STARTTLS, required. Not a hardening extra: SmtpProviderProperties.TlsMode has two members and
+      # neither is plaintext, so the platform cannot describe an unencrypted relay at all. A lane that
+      # wanted a plaintext Mailpit would be asking for a transport the type refuses to express, and
+      # the honest way to satisfy it is to give the relay a certificate.
+      MP_SMTP_TLS_CERT: /run/mailpit-tls/server.crt
+      MP_SMTP_TLS_KEY: /run/mailpit-tls/server.key
+      MP_SMTP_REQUIRE_STARTTLS: "true"
+    volumes:
+      # Generated per run by the qualification wrapper for the host name `mailpit`, and removed on
+      # teardown, exactly like the PostgreSQL lane certificate. A committed test certificate is a
+      # private key in Git.
+      - type: bind
+        source: ./infra/mailpit/tls
+        target: /run/mailpit-tls
+        read_only: true
+    networks:
+      - caskeleton-infra
+    healthcheck:
+      test: ["CMD", "/mailpit", "readyz"]
+      interval: 5s
+      timeout: 3s
+      retries: 12
+      start_period: 5s
+
+  # ---- MinIO -------------------------------------------------------------------
+  minio:
+    profiles:
+      - shared-infra
+      - prod-smoke
+    image: minio/minio:RELEASE.2024-09-13T20-26-02Z
+    command: ["server", "/data"]
+    environment:
+      MINIO_ROOT_USER: "${MINIO_ROOT_USER:-caskeleton}"
+      MINIO_ROOT_PASSWORD: "${MINIO_ROOT_PASSWORD:-caskeleton-local}"
+    volumes:
+      - type: volume
+        source: caskeleton-minio-data
+        target: /data
+    networks:
+      - caskeleton-infra
+    healthcheck:
+      test: ["CMD", "mc", "ready", "local"]
+      interval: 10s
+      timeout: 5s
+      retries: 10
+      start_period: 10s
+
+  # Bucket and policy bootstrap. Not a substitute for the round trip: creating a bucket proves the
+  # server accepts an admin command, not that an object survives being written and read back.
+  minio-init:
+    profiles:
+      - shared-infra
+      - prod-smoke
+    image: minio/mc:RELEASE.2024-09-16T17-43-14Z
+    depends_on:
+      minio:
+        condition: service_healthy
+    entrypoint: ["/bin/sh", "/opt/minio/bucket-bootstrap.sh"]
+    environment:
+      MINIO_ENDPOINT: "http://minio:9000"
+      MINIO_ROOT_USER: "${MINIO_ROOT_USER:-caskeleton}"
+      MINIO_ROOT_PASSWORD: "${MINIO_ROOT_PASSWORD:-caskeleton-local}"
+      MINIO_BUCKET: "${MINIO_BUCKET:-ca-skeleton-objects}"
+    volumes:
+      - type: bind
+        source: ./infra/minio/init
+        target: /opt/minio
+        read_only: true
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+  # ---- Keycloak ----------------------------------------------------------------
+  keycloak:
+    profiles:
+      - local-graphql
+      - local-notification-ingest
+      - local-notification-serving
+      - local-notification-handoff
+      - shared-infra
+      - prod-smoke
+      - all-adapters
+    image: quay.io/keycloak/keycloak:26.0
+    # The wrapper reads the client secret from a mounted file and execs kc.sh. The realm artifact
+    # carries only a ${...} reference, so no secret value is in Git, in the rendered config, or on a
+    # command line.
+    entrypoint: ["/bin/bash", "/opt/keycloak-entrypoint/entrypoint.sh"]
+    environment:
+      KC_BOOTSTRAP_ADMIN_USERNAME: "${KEYCLOAK_ADMIN:-admin}"
+      KC_BOOTSTRAP_ADMIN_PASSWORD: "${KEYCLOAK_ADMIN_PASSWORD:-admin}"
+      KC_HEALTH_ENABLED: "true"
+    volumes:
+      - type: bind
+        source: ./infra/keycloak/entrypoint.sh
+        target: /opt/keycloak-entrypoint/entrypoint.sh
+        read_only: true
+      - type: bind
+        source: ./infra/keycloak/realms
+        target: /opt/keycloak/data/import
+        read_only: true
+    secrets:
+      - keycloak-graphql-smoke-client-secret
+    networks:
+      - caskeleton-infra
+    healthcheck:
+      test:
+        - "CMD-SHELL"
+        - "exec 3<>/dev/tcp/127.0.0.1/9000 && echo -e 'GET /health/ready HTTP/1.1\\r\\nHost: localhost\\r\\nConnection: close\\r\\n\\r\\n' >&3 && cat <&3 | grep -q '\"status\": \"UP\"'"
+      interval: 10s
+      timeout: 5s
+      retries: 20
+      start_period: 30s
+
+  # ---- Capability schema streams ------------------------------------------------
+  # Two pre-start one-shots, in this order, because a capability stream is an operator sequence
+  # rather than a property.
+  #
+  # Install: each stream under db/migration/jpa keeps its own Flyway history table — they all declare
+  # a V1, so one Flyway pointed at all of them fails outright — and each registers itself
+  # INSTALLED_INACTIVE.
+  #
+  # Promote: an operator sanctions the installed schema, and the application refuses to start until
+  # that has happened. That is the fail-closed half of the same design, so it cannot be folded into
+  # the install step without making "the tables exist" and "this is sanctioned" the same event.
+  #
+  # They are also two images because they must be: flyway/flyway ships no psql, so the promotion
+  # could not have run in the migration container at all.
+  #
+  # Both run before `up`, not with the smoke clients after it — the application is what they are a
+  # precondition for. The lane contract's preStartServices carries that ordering.
+  db-migrate-capabilities:
+    profiles:
+      - local-notification-ingest
+      - local-notification-serving
+      - local-notification-handoff
+      - all-adapters
+    image: flyway/flyway:11.1.0
+    depends_on:
+      db:
+        condition: service_healthy
+    entrypoint: ["/bin/sh", "/opt/capability-streams/apply-capability-streams.sh"]
+    environment:
+      PGHOST: "db"
+      PGUSER: "${APP_DATASOURCE_USERNAME:-ca_skeleton}"
+      PGPASSWORD: "${APP_DATASOURCE_PASSWORD:-ca_skeleton}"
+      PGDATABASE: "${POSTGRES_DB:-ca_skeleton}"
+      CAPABILITY_STREAMS: "${CAPABILITY_STREAMS:-}"
+    volumes:
+      - type: bind
+        source: ./infra/postgres/apply-capability-streams.sh
+        target: /opt/capability-streams/apply-capability-streams.sh
+        read_only: true
+      # The whole migration tree, not just db/migration/jpa: the application's own postgresql stream
+      # has to be installed first, or the capability tables arrive in a schema whose flyway_schema_history
+      # does not exist yet and the application refuses to start — which is its baseline-on-migrate: false
+      # policy working as designed.
+      - type: bind
+        source: ./src/adapter/outbound/persistence-jpa/src/main/resources/db/migration
+        target: /flyway/sql
+        read_only: true
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+  db-promote-capabilities:
+    profiles:
+      - local-notification-ingest
+      - local-notification-serving
+      - local-notification-handoff
+      - all-adapters
+    image: postgres:16-alpine
+    depends_on:
+      db:
+        condition: service_healthy
+    entrypoint: ["/bin/sh", "/opt/capability-streams/promote-capability-streams.sh"]
+    environment:
+      PGHOST: "db"
+      PGUSER: "${APP_DATASOURCE_USERNAME:-ca_skeleton}"
+      PGPASSWORD: "${APP_DATASOURCE_PASSWORD:-ca_skeleton}"
+      PGDATABASE: "${POSTGRES_DB:-ca_skeleton}"
+      CAPABILITY_STREAMS: "${CAPABILITY_STREAMS:-}"
+    volumes:
+      - type: bind
+        source: ./infra/postgres/promote-capability-streams.sh
+        target: /opt/capability-streams/promote-capability-streams.sh
+        read_only: true
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+  # The GraphQL transport, as a request. auth-smoke proves a token can be obtained and that public
+  # health answers; this proves /graphql is guarded and that an authenticated query executes.
+  graphql-smoke:
+    profiles:
+      - local-graphql
+      - all-adapters
+    image: curlimages/curl:8.10.1
+    depends_on:
+      keycloak:
+        condition: service_healthy
+    # uid 0 for the mounted 0600 client secret, same as auth-smoke.
+    user: "0:0"
+    entrypoint: ["/bin/sh", "/opt/graphql-smoke/graphql-smoke.sh"]
+    environment:
+      APP_BASE_URL: "http://app:8080"
+      KEYCLOAK_ISSUER: "http://keycloak:8080/realms/ca-skeleton"
+      KEYCLOAK_CLIENT_ID: "ca-skeleton-api"
+      # Spring for GraphQL serves its own endpoint through a router function rather than an
+      # annotated controller, so the presentation base-path prefix does not apply to it.
+      GRAPHQL_PATH: "${GRAPHQL_PATH:-/graphql}"
+    volumes:
+      - type: bind
+        source: ./infra/graphql/smoke
+        target: /opt/graphql-smoke
+        read_only: true
+    secrets:
+      - keycloak-graphql-smoke-client-secret
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+  # ---- One-shot smoke clients --------------------------------------------------
+  # Never `up --wait` targets. Each is run with `run --rm` and must exit zero; a missing, skipped or
+  # non-zero one fails its lane rather than being treated as "not applicable".
+  auth-smoke:
+    profiles:
+      - local-graphql
+      - shared-infra
+      - prod-smoke
+      - all-adapters
+    image: curlimages/curl:8.10.1
+    depends_on:
+      keycloak:
+        condition: service_healthy
+    # The client secret is written on the host at mode 0600 by the qualification wrapper and mounted
+    # in. The Keycloak image happens to run as the same uid the wrapper writes as; this image runs as
+    # uid 100, so it read "Permission denied" and the lane failed on the smoke client rather than on
+    # anything it was checking. Compose ignores the secret's uid/gid/mode options outside swarm, so
+    # the container reads it as root instead. The two alternatives are both worse: loosening the host
+    # file to world-readable leaves a credential readable by every process on the machine, and passing
+    # the value as an environment variable puts it in `docker compose config` output and in ps.
+    user: "0:0"
+    entrypoint: ["/bin/sh", "/opt/auth-smoke/auth-smoke.sh"]
+    environment:
+      # The same issuer URL the application is given. A token obtained from one URL and validated
+      # against another proves nothing, and localhost means a different host inside each container.
+      KEYCLOAK_ISSUER: "http://keycloak:8080/realms/ca-skeleton"
+      KEYCLOAK_CLIENT_ID: "ca-skeleton-api"
+      APP_BASE_URL: "http://app:8080"
+      # Supplied per runtime, because the same endpoint has two addresses: application-local.yml
+      # pins presentation.api-base-path to /api and the shipped default is /v1. The qualification
+      # wrapper exports the value that matches the lane's Spring runtime.
+      APP_HEALTH_PATH: "${APP_HEALTH_PATH:-/v1/healthcheck}"
+    volumes:
+      - type: bind
+        source: ./infra/keycloak/smoke
+        target: /opt/auth-smoke
+        read_only: true
+    secrets:
+      - keycloak-graphql-smoke-client-secret
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+  # The server image, not the mc client image: minio/mc ships no sed, grep or cmp, and the round-trip
+  # client needs a digest tool. See infra/minio/smoke/object-storage-smoke.sh for how that went
+  # unnoticed. The lane already pulls this image for the server itself.
+  object-storage-smoke:
+    profiles:
+      - shared-infra
+      - prod-smoke
+    image: minio/minio:RELEASE.2024-09-13T20-26-02Z
+    depends_on:
+      minio-init:
+        condition: service_completed_successfully
+    entrypoint: ["/bin/sh", "/opt/minio-smoke/object-storage-smoke.sh"]
+    environment:
+      MINIO_ENDPOINT: "http://minio:9000"
+      MINIO_ROOT_USER: "${MINIO_ROOT_USER:-caskeleton}"
+      MINIO_ROOT_PASSWORD: "${MINIO_ROOT_PASSWORD:-caskeleton-local}"
+      MINIO_BUCKET: "${MINIO_BUCKET:-ca-skeleton-objects}"
+    volumes:
+      - type: bind
+        source: ./infra/minio/smoke
+        target: /opt/minio-smoke
+        read_only: true
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+  notification-smoke:
+    profiles:
+      - local-notification-ingest
+      - local-notification-serving
+      - local-notification-handoff
+      - all-adapters
+    image: curlimages/curl:8.10.1
+    depends_on:
+      keycloak:
+        condition: service_healthy
+    # uid 0 for the same reason auth-smoke uses it: the mounted client secret is mode 0600 on the
+    # host and this image otherwise runs as uid 100, which reads "Permission denied". The lane then
+    # fails on the smoke client rather than on anything it was checking.
+    user: "0:0"
+    entrypoint: ["/bin/sh", "/opt/notification-smoke/notification-smoke.sh"]
+    environment:
+      APP_BASE_URL: "http://app:8080"
+      MAILPIT_BASE_URL: "http://mailpit:8025"
+      # ingest | serving | handoff-verify — which phase of the lane this invocation is.
+      #
+      # No default, deliberately. It defaulted to `ingest`, and local-notification-serving therefore
+      # ran the ingest assertions — "accepted, and nothing was delivered" — against an application in
+      # SERVING mode. The lane passed while testing the opposite of what it is named for, and would
+      # have kept passing for as long as the check happened to run before the dispatch worker. An
+      # unset value now renders empty and the client refuses it.
+      # `:-` and not a value: an explicit empty default keeps Compose from warning about an unset
+      # variable on every lane that never runs this client, while still rendering empty so the
+      # client refuses it.
+      NOTIFICATION_SMOKE_PHASE: "${NOTIFICATION_SMOKE_PHASE:-}"
+      # Submission and template publication are authenticated like every other non-public path, so
+      # this client obtains a token the same way auth-smoke does — client credentials against the
+      # same issuer URL the application validates against.
+      APP_BASE_PATH: "${APP_BASE_PATH:-/api}"
+      KEYCLOAK_ISSUER: "http://keycloak:8080/realms/ca-skeleton"
+      KEYCLOAK_CLIENT_ID: "ca-skeleton-api"
+    volumes:
+      - type: bind
+        source: ./infra/notification/smoke
+        target: /opt/notification-smoke
+        read_only: true
+      # The handoff lane runs this client twice in one project and the second run needs the request
+      # id the first accepted, so the state lives in a named volume that outlives a `run --rm`
+      # container and is removed with the project by the teardown's --volumes.
+      #
+      # Its own path, not a subdirectory of the script mount above: a volume nested inside a
+      # read-only bind cannot be created, because the runtime has to mkdir the mountpoint in a
+      # filesystem it was just told is read-only.
+      - type: volume
+        source: caskeleton-notification-smoke-state
+        target: /opt/notification-smoke-state
+    secrets:
+      - keycloak-graphql-smoke-client-secret
+    networks:
+      - caskeleton-infra
+    restart: "no"
+
+networks:
+  caskeleton-infra:
+    driver: bridge
+
+volumes:
+  caskeleton-notification-smoke-state:
+    driver: local
+  caskeleton-db-data:
+    driver: local
+  caskeleton-mongo-data:
+    driver: local
+  caskeleton-minio-data:
+    driver: local
+
+secrets:
+  # Written per run at mode 0600 by the qualification wrapper and removed on teardown. The realm
+  # artifact references it by name; the value never reaches Git, a rendered config, a command line,
+  # or an evidence file.
+  keycloak-graphql-smoke-client-secret:
+    file: ./infra/keycloak/secrets/graphql-smoke-client-secret
@@ -5,19 +5,32 @@
 #   docker compose -f docker-compose.yml -f docker-compose.local.yml up
 #
 # Local intent:
-#   - Starts a local PostgreSQL database for integration testing without Testcontainers.
-#   - Wires the app environment to point at the local DB.
+#   - Wires the app environment to point at the shared `db` service, which lives in
+#     docker-compose.infra.yml and starts only for lanes whose Compose profile names it.
+#   - Declares no depends_on: a depends_on aimed at a profiled service makes every lane that does
+#     not enable that profile fail to render at all, and ordering is the runtime-smoke wrapper's
+#     job — it knows which services a lane actually starts.
 #   - Keeps read-only filesystem and memory limits from the base compose.
-#   - Does NOT expose the DB port publicly; app and db communicate on the
-#     internal `caskeleton-local` network only.
+#   - Publishes the DB on the loopback interface only, so a host-side run
+#     (`./gradlew :app-bootstrap:bootRun`, IDE) reaches the same database the
+#     containerised app reaches over the internal `caskeleton-local` network.
 # =============================================================================

 services:
  app:
+    # Optional, because src/.env is operator input and a fresh clone does not have one. Before this
+    # was marked optional, untracking that file made `docker compose config` fail outright on a
+    # clone — the environment override that exists for convenience became a hard prerequisite for
+    # rendering the stack at all. The tracked contract is src/.env.example; copy it.
    env_file:
-      - ./src/.env
+      - path: ./src/.env
+        required: false
    # Wire the app to the local Postgres service on the internal network.
    environment:
+      # Explicit, not inherited. A Compose profile selects services; it says nothing about which
+      # environment the application believes it is in, and the two drifting is how a dev stack ends
+      # up running local's settings.
+      SPRING_PROFILES_ACTIVE: "local"
      TZ: "UTC"
      LANG: "C.UTF-8"
      LC_ALL: "C.UTF-8"
@@ -29,9 +42,6 @@ services:
      APP_DATASOURCE_URL: "jdbc:postgresql://db:5432/${POSTGRES_DB:-ca_skeleton}"
      APP_DATASOURCE_USERNAME: "${APP_DATASOURCE_USERNAME:-ca_skeleton}"
      APP_DATASOURCE_PASSWORD: "${APP_DATASOURCE_PASSWORD:-ca_skeleton}"
-    depends_on:
-      db:
-        condition: service_healthy
    healthcheck:
      test:
        - "CMD"
@@ -45,39 +55,10 @@ services:
      start_period: 20s
      retries: 12
    networks:
-      - caskeleton-local
-
-  db:
-    image: postgres:16-alpine
-    environment:
-      POSTGRES_DB: "${POSTGRES_DB:-ca_skeleton}"
-      POSTGRES_USER: "${APP_DATASOURCE_USERNAME:-ca_skeleton}"
-      POSTGRES_PASSWORD: "${APP_DATASOURCE_PASSWORD:-ca_skeleton}"
-      TZ: "UTC"
-    # Persist data between restarts; remove the volume to start fresh.
-    volumes:
-      - type: volume
-        source: caskeleton-db-data
-        target: /var/lib/postgresql/data
-    # No host port: startup Flyway runs in the app container over the internal network.
-    networks:
-      - caskeleton-local
-    healthcheck:
-      test:
-        [
-          "CMD-SHELL",
-          "pg_isready -U ${APP_DATASOURCE_USERNAME:-ca_skeleton} -d ${POSTGRES_DB:-ca_skeleton}",
-        ]
-      interval: 10s
-      timeout: 5s
-      retries: 5
-      start_period: 30s
-    restart: unless-stopped
+      - caskeleton-infra

 networks:
-  caskeleton-local:
-    driver: bridge
+  # Defined in docker-compose.infra.yml, where the services that share it live.
+  caskeleton-infra:
+    external: false

-volumes:
-  caskeleton-db-data:
-    driver: local
@@ -0,0 +1,37 @@
+# =============================================================================
+# prod-smoke — a production-shaped runtime, for evidence, on a laptop.
+#
+# Not "production Compose". What it is for is proving that the prod profile's fail-closed validators
+# can be satisfied at all: TLS on the JDBC URL, a schema Flyway owns, JSON logging, secret
+# references rather than values. A prod lane that only ever gets as far as `config` proves the file
+# parses, which was never the thing in doubt.
+#
+# The credentials here are generated per run by the lane wrapper. Nothing local is reused: a
+# prod-smoke that borrows the local MinIO password is a prod-smoke that tests the local setup.
+# =============================================================================
+
+services:
+  app:
+    environment:
+      # Explicit, not inherited. A Compose profile selects services and says nothing about which
+      # environment the application believes it is in.
+      SPRING_PROFILES_ACTIVE: "prod"
+      TZ: "UTC"
+      # verify-full, which is the point: PostgreSqlTransportSecurityValidator refuses anything less,
+      # and that refusal is the behaviour this lane exists to satisfy rather than bypass.
+      APP_DATASOURCE_URL: "jdbc:postgresql://db:5432/${POSTGRES_DB:-ca_skeleton}?sslmode=verify-full&sslrootcert=/run/secrets/postgres-ca"
+      APP_DATASOURCE_USERNAME: "${APP_DATASOURCE_USERNAME:-ca_skeleton}"
+      # The password is deliberately absent here. An `environment:` entry beats `env_file:`, so
+      # declaring it as "${APP_DATASOURCE_PASSWORD:-}" read the host shell rather than the lane's
+      # generated file and injected an empty string — which the prod env validator then refused, for
+      # the right reason, about a value the lane had actually supplied.
+      APP_DATASOURCE_DDL_AUTO: "validate"
+      APP_LOG_JSON_ENABLED: "true"
+      APP_SECURITY_JWT_ISSUER: "http://keycloak:8080/realms/ca-skeleton"
+      APP_SECURITY_JWT_AUDIENCE: "ca-skeleton-api"
+    networks:
+      - caskeleton-infra
+
+networks:
+  caskeleton-infra:
+    external: false
@@ -0,0 +1,50 @@
+# =============================================================================
+# Database transport security, for the lanes whose runtime requires it.
+#
+# The prod runtime connects with `sslmode=verify-full` and an explicit `sslrootcert`. That is not a
+# lane setting to relax: a prod smoke test against a database with TLS disabled is a smoke test of a
+# configuration production never runs, and the one failure mode it would hide — the certificate
+# chain or the host name not checking out — is the one that only ever appears in production.
+#
+# So the lane brings a real certificate instead. The qualification wrapper generates a CA and a
+# server certificate for the host name `db` per run, at mode 0600, and removes both on teardown; the
+# realm-secret pattern, applied to a keypair. Nothing here is committed: infra/postgres/tls holds
+# only a .gitignore.
+#
+# `verify-full` is deliberate rather than `verify-ca`. `verify-ca` proves the certificate was issued
+# by the expected authority and says nothing about who presented it, so it does not detect a
+# redirected connection — which is most of what transport security is for.
+# =============================================================================
+
+services:
+  db:
+    # Runs as root just long enough to install the key where postgres can read it, then hands over
+    # to the official entrypoint. See infra/postgres/entrypoint.sh for why a bind mount cannot do it.
+    entrypoint: ["/bin/sh", "/opt/postgres-entrypoint/entrypoint.sh"]
+    command:
+      - "postgres"
+      - "-c"
+      - "ssl=on"
+      - "-c"
+      - "ssl_cert_file=/etc/postgresql-tls/server.crt"
+      - "-c"
+      - "ssl_key_file=/etc/postgresql-tls/server.key"
+    volumes:
+      - type: bind
+        source: ./infra/postgres/entrypoint.sh
+        target: /opt/postgres-entrypoint/entrypoint.sh
+        read_only: true
+      - type: bind
+        source: ./infra/postgres/tls
+        target: /opt/postgres-tls
+        read_only: true
+
+  app:
+    # The certificate authority the JDBC URL names in `sslrootcert`. A public certificate, so it
+    # carries no mode problem — the private half never leaves the database container's filesystem.
+    secrets:
+      - postgres-ca
+
+secrets:
+  postgres-ca:
+    file: ./infra/postgres/tls/ca.crt
@@ -34,6 +34,14 @@ services:
        GIT_SHA: "${GIT_SHA:-0000000}"
        SOURCE_URL: "${SOURCE_URL:-https://example.invalid/ca-tmpl}"
    image: caskeleton:${BUILD_VERSION:-0.0.1_local_0000000}
+    # Generated per run by scripts/run-compose-runtime-smoke.sh and removed on teardown. Seven values
+    # have no inline default on purpose — the datasource address and credential, the application
+    # name, and the JWT issuer and audience — so a lane has to supply them, and a lane that borrowed
+    # the developer's own src/.env would be reproducible only on that developer's machine. Optional,
+    # so an ordinary `docker compose up` is unaffected.
+    env_file:
+      - path: ./src/.env.lane
+        required: false
    ports:
      - "${APP_SERVER_PORT:-8080}:8080"
      - "9001:9001"
@@ -53,6 +61,20 @@ services:
    tmpfs:
      - /tmp:mode=1777,size=128m
      - /var/tmp/heap:mode=1777,size=512m
+    # ---- Fileserver storage volume ------------------------------------------
+    # A named volume, not a tmpfs and not the read-only root. The Fileserver platform's default
+    # storage root is /var/lib/backend/files, and with a read-only root and no mount there was
+    # nowhere on the image it could legally write: enabling the capability failed on its first
+    # upload rather than at startup. The volume is declared unconditionally because a volume
+    # nobody writes to costs nothing, while a missing one costs an outage.
+    #
+    # Ownership: the image runs as uid/gid 1000 (see src/Dockerfile). Docker initialises a fresh
+    # named volume from the image path's ownership, so the directory is created in the image with
+    # that owner; a pre-existing volume or a host bind mount must be chowned to 1000:1000 by the
+    # operator, or every write is refused with a permission error the application reports as
+    # STORAGE_UNAVAILABLE.
+    volumes:
+      - fileserver-data:/var/lib/backend/files
    # ---- Memory limit (D4) --------------------------------------------------
    # Must be set so -XX:MaxRAMPercentage=75 can compute a meaningful heap bound.
    mem_limit: 512m
@@ -78,3 +100,8 @@ services:
      start_period: 60s
      retries: 3
    restart: unless-stopped
+
+volumes:
+  # Survives container replacement, which is the point: published content outlives the process
+  # that wrote it. Back this with real storage in any deployment that keeps files.
+  fileserver-data:
@@ -0,0 +1,54 @@
+# docs
+
+저장소의 모든 문서는 이 디렉터리 아래에 있다. 어떤 문서를 어디에 두는지가 유일한 규칙이고,
+파일 목록은 디렉터리를 직접 읽는다. 개수를 여기에 적으면 다음 문서가 추가되는 순간 틀린 글이 된다.
+
+## 어댑터별 운영 문서
+
+각 어댑터의 지원 범위, 설정, 보안, 운영, 마이그레이션 문서다. 코드와 함께 갱신되어야 하는 문서이고,
+`docs/httpclient/` 는 `scripts/verify-httpclient-docs.py` 가 코드에서 뽑은 이름과 대조한다.
+
+| 디렉터리 | 대상 |
+| --- | --- |
+| `fileserver/` | 파일 서버 어댑터 |
+| `httpclient/` | HTTP 클라이언트 플랫폼 |
+| `jpa/` | JPA·PostgreSQL 영속성 |
+| `messaging/` | 메시징 어댑터 |
+| `mongodb/` | MongoDB 문서 영속성 (`advanced/`, `runbooks/` 포함) |
+| `notification/` | 알림 전달 플랫폼 (`adr/` 포함) |
+| `redis/` | Redis 캐시·세션 |
+
+## 횡단 문서
+
+| 디렉터리 | 대상 |
+| --- | --- |
+| `adr/` | 아키텍처 결정 기록 |
+| `architecture/` | 공개 API 표면 스냅숏 |
+| `evidence/` | 작업 단계별 증거·체크포인트 |
+| `registries/` | env 키·에러 코드·메트릭·헤더 등 레지스트리 SSOT |
+| `reviews/` | 모듈 코드 리뷰 결과 |
+| `runbooks/` | 장애 코드별 대응 런북 (`template.md` 기준) |
+| `security/` | 공개 경로 스냅숏 |
+
+## 설계와 계획
+
+| 디렉터리 | 대상 |
+| --- | --- |
+| `superpowers/specs/` | 설계서. `YYYY-MM-DD-<주제>-design.md` |
+| `superpowers/plans/` | 구현·확장 계획서. `YYYY-MM-DD-<주제>-plan.md` |
+| `superpowers/packages/` | 외부에서 납품된 설계 패키지의 README와 정적 검증 결과 |
+
+`superpowers/packages/<어댑터>/` 는 설계서가 처음 전달됐을 때의 안내와 `VALIDATION.md` 검증 이력을
+남긴 기록 보관소다. 설계서·계획서 본문은 전부 `specs/` 와 `plans/` 에 있으므로 이 디렉터리에서
+문서를 찾을 필요는 없다. 각 README 상단의 보존 안내가 무엇이 옮겨졌고 무엇이 제거됐는지 밝힌다.
+
+계획서 본문에는 당시 계획한 경로와 명령이 그대로 남아 있다. 그중 일부는 실제 구현에서 다른 위치로
+조정됐고, 저장소에 어떻게 대응시켰는지는 각 어댑터의 `repository-adaptation.md` 또는
+`module-mapping.md` 가 기록한다. 계획서를 사후에 고치지 않는 이유는 그렇게 하면 계획의 기록이 아니라
+결과를 계획처럼 보이게 만든 글이 되기 때문이다.
+
+## 여기에 없는 것
+
+- 실행되는 검증 스크립트는 문서가 아니다. `scripts/` 와 `.github/scripts/` 에 있다.
+- 모듈 레지스트리·Gradle 정책은 `src/config/architecture/modules.json` 과 `src/build.gradle` 이 소유한다.
+- 각 모듈의 지역 규칙은 해당 모듈의 `src/**/CLAUDE.md` 가 소유한다.
@@ -0,0 +1,119 @@
+# ADR-BUILD-001: `java-test-fixtures` is the standard for shared test code
+
+- Status: Accepted
+- Date: 2026-09-07
+- Scope: every leaf that publishes or consumes shared test code
+- Source: `docs/reviews/2026-09-07-app-bootstrap-module-code-review.md` BOOT-015
+
+## Context
+
+Two conventions do the same job in this repository.
+
+`ca.testkit-publisher` — a convention plugin — gives a leaf a `testkit` source set, wires its output
+onto the lanes that leaf names, and optionally publishes it as a consumable configuration. Five
+leaves use it: `persistence-jpa` (published as `jpaTestkit`), `web` (`webTestkit`), `websocket`
+(`websocketTestkit`), `persistence-mongo` and `httpclient` (both unpublished).
+
+`java-test-fixtures` — Gradle's own plugin — gives a leaf a `testFixtures` source set, puts it on
+`test`'s classpath automatically, and always publishes it as a variant consumers reach with
+`testFixtures(project(':x'))`. One leaf uses it: `graphql`, which additionally fails its build when a
+fixture is written outside `src/testFixtures/java`.
+
+Two conventions for one purpose is the defect. A contributor adding shared test code has to know
+which leaf they are in before they know where the file goes, and the two answers are not
+interchangeable: a consumer of the first writes `project(path: ':x', configuration: 'jpaTestkit')`
+and has to know the configuration's name, while a consumer of the second writes
+`testFixtures(project(':x'))` and does not.
+
+## Decision
+
+**`java-test-fixtures` is the standard.** New shared test code goes in `src/testFixtures/java`, and a
+consumer depends on it with `testFixtures(project(':x'))`.
+
+Three reasons, in order of weight:
+
+1. **The consumer side describes itself.** `testFixtures(project(':adapter:inbound:web'))` says what
+   it is. `project(path: ':adapter:inbound:web', configuration: 'webTestkit')` says where to look,
+   and only after the reader has learned that `webTestkit` is a testkit rather than a lane.
+2. **The enforcement already exists and is copyable.** `graphql`'s build fails when a fixture is
+   declared in the wrong place. The same guard applies unchanged to any leaf that adopts the plugin.
+3. **It is one fewer local concept.** A convention plugin that reimplements a Gradle plugin has to be
+   maintained against it.
+
+## What the local plugin does better, and how it is replaced
+
+This is worth writing down, because the review that prompted this ADR recommended the migration
+before reading `ca.testkit-publisher`, and the plugin turns out to encode two deliberate decisions
+rather than being an oversight.
+
+**Publishing is opt-in.** `persistence-mongo` and `httpclient` have a testkit and publish nothing;
+`persistence-jpa` publishes. The plugin's own comment names this as "a real difference in what each
+leaf offers rather than an oversight to normalise away". `java-test-fixtures` always creates the
+variant, so the distinction is lost — a leaf that never meant to offer its fixtures will offer them.
+
+> Replacement: none at the build level. The distinction moves to review: the fixtures of a leaf that
+> nobody consumes are simply unconsumed. This is a real, accepted loss.
+
+**Lane consumption is declared.** `persistence-jpa` says `consumedBy 'test', 'postgresqlIntegrationTest'`.
+`java-test-fixtures` puts fixtures on `test` only, so every other lane needs the output added
+explicitly.
+
+> Replacement: `strictTestLanes`' existing `compilesAgainst` expresses this unchanged — a lane
+> declares `compilesAgainst 'main', 'testFixtures'`. The first draft of this ADR assumed the DSL
+> would need a change, because `sourceSet(name)` creates what it is given and `testFixtures` already
+> exists. The `persistence-mongo` migration showed otherwise: `compilesAgainst` only *looks a source
+> set up*, so naming a plugin-created one works as-is. What the leaf drops is the
+> `sourceSet('testkit')` declaration, not the lane's.
+
+## Migration: done, and what it cost
+
+Five leaves, eleven lanes, two published testkits, all migrated leaf by leaf with the suite run
+between each. `ca.testkit-publisher` is deleted.
+
+The order was chosen so a mistake would be cheap: unpublished leaves first, published ones last with
+their consumer in the same step.
+
+1. `persistence-mongo` — one leaf, two lanes, no cross-module consumer; the proof the path works.
+   What it took, per leaf:
+   - `apply plugin: 'java-test-fixtures'` at the top of the leaf build file;
+   - `git mv src/testkit src/testFixtures`;
+   - drop `sourceSet('testkit')` and the whole `testkitPublisher` block; keep every other lane's
+     `compilesAgainst`, renaming `'testkit'` to `'testFixtures'`;
+   - rename `testkitImplementation` to `testFixturesImplementation`, **and add what the old source
+     set was inheriting silently**. This is the one non-mechanical step: `testkit*` extended
+     `testImplementation`, so the fixtures saw every test library the leaf declared. Mongo's needed
+     four more lines (AssertJ, BSON, Spring Data commons, Toxiproxy) — none of which the leaf had
+     ever stated the fixtures depended on;
+   - regenerate the leaf's lock state.
+2. `httpclient`, then `websocket` — unpublished as well, more lanes.
+3. `web` and `persistence-jpa` with `app-bootstrap`'s two consumer declarations, which became
+   `testImplementation(testFixtures(project(':…')))`.
+4. `ca.testkit-publisher` deleted, along with its `plugins {}` entry and its application in the root
+   build.
+
+### Two things the migration broke, and what they taught
+
+Both were caught by tests that exist to catch exactly this, which is the argument for having them.
+
+**ArchUnit corpora went wrong in opposite directions.** `httpclient`'s boundary rules *excluded*
+`build/classes/java/testkit`; after the move the fixtures arrived as a `…-test-fixtures.jar` on the
+same classpath, so the exclusion missed them and 258 fixture-to-fixture calls were reported as
+production depending on the testkit. `persistence-jpa`'s rules *included* only
+`build/classes/java/main`; applying `java-test-fixtures` makes the module's own test classpath carry
+the module as a **jar** rather than as a class directory, so its corpus became empty. The second is
+the dangerous one — an empty corpus makes every `noClasses()` rule pass — and it surfaced only
+because that suite asserts its corpus is non-empty before asserting anything about it.
+
+**Fixtures had invisible dependencies.** `testkit*` configurations extended `testImplementation`, so
+the fixtures compiled against every test library their leaf declared without ever naming one. Making
+them explicit took roughly thirty `testFixturesImplementation` lines across the five leaves —
+Micrometer, Spring Web, Netty, logback, Jackson, JUnit, AssertJ, Spring Data. None of them were
+wrong; none of them were stated.
+
+## Consequences
+
+- `docs/testing/TESTING_STRATEGY.md` §5 records the standard; this ADR records why and at what cost.
+- Until step 5, two conventions remain visible. The strategy document says so explicitly, so a
+  contributor reading it is not left to infer which one is current.
+- The opt-in-publishing distinction is given up. If it later proves load-bearing — a leaf whose
+  fixtures genuinely must not be reachable — the answer is a separate module, not a third convention.
@@ -0,0 +1,70 @@
+# ADR-GQL-001 — GraphQL context stays inbound; object authorization moves to application-core; the persisted-operation store stays an inbound SPI
+
+- Status: Accepted
+- Date: 2026-08-24
+- Review: `docs/reviews/2026-08-14-graphql-module-code-review.md` GQL-026
+
+## Context
+
+The GraphQL leaf's own documentation described three things crossing its boundary: a
+`GraphQlRequestContext` with a deadline propagated into application, JPA, Mongo and the HTTP client;
+object authorization decided inside the transport; and a persisted-operation registry implemented by
+an external durable store.
+
+Two of those invert the dependency direction. If `application-core` or an outbound adapter
+implements a type that lives in `adapter:inbound:graphql`, the registry edge that says inbound
+depends on application is satisfied while the real compile-time dependency runs the other way.
+
+The third is a business rule in the wrong layer: whether an actor may see an object is a decision
+about the domain, and GraphQL is one of four transports this skeleton ships.
+
+## Decision
+
+Three different answers, because the three problems are not the same problem.
+
+**GraphQL context stays inbound-local.** It is mapped explicitly onto application command fields —
+actor, tenant, deadline — rather than travelling as a type. Nothing outside the leaf references
+`GraphQlRequestContext`, and the boundary test is that grep returns nothing outside it.
+
+**Object authorization moves to `application-core`.** `ObjectAccessPolicy`, `ObjectAccessRequest`
+and `ObjectAccessDecision` are transport-neutral and live with the other application policies;
+`ApplicationObjectAuthorization` in the GraphQL leaf is the bridge that calls them. This is the one
+of the three that was a real layering defect, and it is fixed rather than documented.
+
+**The persisted-operation store stays an inbound-owned SPI.** `GraphQlPersistedOperationRegistry`
+remains in `advanced/persisted`, and no leaf outside GraphQL implements it.
+
+## Consequences
+
+The third decision is the one that needs defending, because it leaves the reported risk in place
+rather than removing it.
+
+The risk is conditional: the direction inverts only when something outside the leaf implements the
+interface. Nothing does. The template ships an in-memory registry and no durable one, because it
+ships no persisted-operation store at all.
+
+The alternative was to introduce a generic operational key-value store port owned by a neutral
+contract holder, with the GraphQL adapter owning only the key and value mapping. That port would
+have exactly one interface, zero implementations and one speculative consumer — a new abstraction
+whose shape is guessed from a requirement nobody has stated. This repository has spent a full
+remediation pass deleting controls that existed and were reached by nothing, and inventing a port
+for a store that does not exist is how the next one of those gets written.
+
+So the decision is to leave the SPI where it is and to move it when a durable store is actually
+built. Moving it then is a rename across one leaf and one new adapter, which is cheaper than
+carrying a wrong abstraction until then. What must not happen in the meantime is an outbound leaf
+implementing the inbound interface, because that is the moment the direction actually inverts, and
+it would happen in a commit whose diff looks like an implementation rather than a layering change.
+
+The composition root wires these and owns no business or storage policy of its own.
+
+## Enforcement
+
+`verifyCleanArchitectureDependencies` and `modules.json` hold the leaf's edges to
+`domain-core`, `application-core` and `shared-contract`. `ObjectAccessPolicyTest` covers the
+application-side policy and `ApplicationObjectAuthorizationTest` the bridge.
+
+The condition this ADR turns on — that nothing outside the GraphQL leaf implements the
+persisted-operation SPI — is a claim about the whole repository, so it is checked at the
+composition root rather than inside the leaf, next to the other GraphQL boundary rules in
+`app-bootstrap`'s architecture suite.
@@ -0,0 +1,55 @@
+# ADR-GRPC-001: The gRPC platform ships as a registered family, not as one adapter leaf
+
+- Status: accepted
+- Date: 2026-08-30
+- Scope: `:grpc:*`, `:grpc-advanced:*`, `src/config/architecture/modules.json`
+
+## Context
+
+The two source plans describe a type-safe gRPC execution platform with its own API, SPI, adapters and
+composition root: fifteen Stable modules under `modules/grpc` and sixteen Advanced ones under
+`modules/grpc-advanced`, on Gradle Kotlin DSL, in package `io.backend.skeleton.grpc`, against
+Spring Boot 4.1.
+
+None of that layout exists here. This repository uses Groovy DSL, a fail-closed module registry that
+owns the leaf list, package root `dev.caskeleton`, and Spring Boot 4.0.8. The plans anticipate this:
+their last Global Constraint says that when the repository structure differs, file paths are remapped
+and the public contracts, invariants and test meanings are not changed.
+
+Two shapes were available. Fold the platform into the existing `:adapter:inbound:grpc` leaf as
+packages — which is what the JPA, GraphQL, WebSocket and HTTP platforms did here — or register it as
+a family the way `messaging:*` is registered.
+
+## Decision
+
+Register it as a family: twelve Stable leaves under `src/grpc/` and six Advanced ones under
+`src/grpc-advanced/`.
+
+The deciding property is that this is not a layer of this application. Root `CLAUDE.md` already
+describes `messaging:*` as "a vendored messaging platform: a product with its own API, SPI, adapters
+and composition boundary, not a layer of this application", and the gRPC platform is the same shape
+for the same reason — the application is meant to reach it the way it reaches a library, through an
+application-owned port. The four platforms that became packages are all layers of this application;
+this one is not.
+
+The split between `src/grpc/` and `src/grpc-advanced/` is not organisational. The Stable plan
+requires that the Stable starter's build fail if it reaches an Advanced module, and separate Gradle
+path prefixes make that a `verifyCleanArchitectureDependencies` failure rather than a review note:
+`grpc-spring-boot-starter`'s registry entry names no advanced id, and it cannot acquire one silently.
+
+## Consequences
+
+**The registry grew from 44 leaves to 62.** That is a large registry change, made deliberately and in
+one place. Every new leaf is `runtime_memberships: []`, so nothing ships until a second, explicit
+decision moves it.
+
+**The advanced boundary is checked twice.** Once by the registry at build time, and once by
+`GrpcStableBuildInvariant` at runtime, because a fat jar or a shaded artifact is assembled by
+something the registry never sees.
+
+**Four testkit modules became four test lanes.** The plan's split exists so in-process results cannot
+be mistaken for network results; this repository expresses that with `ca.strict-test-lane`, whose
+lanes fail when they discover nothing and never serve an up-to-date result. `GrpcEvidenceGrade` keeps
+the same rule inside the code, so a report cannot cite a contract run as transport evidence.
+
+**Codegen is not wired.** See ADR-GRPC-002.
@@ -0,0 +1,62 @@
+# ADR-GRPC-002: Schema governance runs without protoc and without the Buf CLI
+
+- Status: accepted
+- Date: 2026-08-30
+- Scope: `:grpc:grpc-proto-contract`, `:grpc:grpc-codegen`
+
+## Context
+
+Stable Tasks 8 through 11 require proto style rules, Buf format/lint/breaking governance, a single
+Java codegen owner, and a descriptor artifact whose consumer-compile result gates a release.
+
+Two of the tools those tasks name are absent from this toolchain. The Buf CLI is not installed. And
+`protoc` is available through the Gradle protobuf plugin, but every leaf in this repository passes
+spotless with google-java-format, checkstyle, SpotBugs at HIGH confidence, Error Prone and `-Werror`
+— and generated protobuf sources pass none of them. Turning codegen on means excluding a source set
+from five quality gates.
+
+There is precedent for such an exclusion: the `jmh` source set has `spotbugsJmh` and `checkstyleJmh`
+disabled and Error Prone off. So the carve-out is available. It is also a decision about the quality
+baseline of a leaf, taken for one task, and outside what this work was asked to change.
+
+`adapter:inbound:grpc` also carries a recorded decision in the opposite direction: its `CLAUDE.md`
+forbids the protobuf plugin and `.proto` in that leaf, on the grounds that a consuming feature module
+should own its schema.
+
+## Decision
+
+Commit the `.proto` sources and implement every rule the tasks require as executable Java, with no
+protoc run and no Buf CLI invocation.
+
+`GrpcProtoContractValidator` reads `.proto` text and enforces proto3 syntax, the
+`{organization}.{domain}.v{major}` package rule, `java_multiple_files`, a generated Java package
+disjoint from the hand-written one, `_UNSPECIFIED` enum zero values, `reserved` declarations checked
+against a supplied removal history, a well-known-type allowlist and a map-field allowlist. It runs
+against the committed schema in its own test, so the shipped `.proto` files are live rather than
+decorative.
+
+`GrpcBufPolicy` fixes the breaking gate at Buf's `FILE` category and names the four lifecycle stages
+a compliant pipeline registers. `GrpcCodegenManifest` fixes one codegen owner and refuses a literal
+generator version. `GrpcDescriptorArtifact`, `GrpcConsumerFixture` and `GrpcSchemaArtifactPublisher`
+carry the schema hash, the descriptor digest and the per-consumer source-break report, and refuse a
+publish that breaks a consumer or republishes a released version with different bytes.
+
+The committed `buf.yaml` states the same rules, so running the CLI in an environment that has it
+reaches the same verdict.
+
+## Consequences
+
+**The invariants are enforced; the process is not run.** Everything Tasks 8 to 11 are about — which
+schema changes are refused, which consumer breaks block a release, who owns generation — is a
+build-checkable rule here. What is missing is the protoc invocation and the Buf binary.
+
+**Turning codegen on is a bounded change.** `GrpcCodegenManifest.caSkeleton()` already names the
+owner, the managed version source, the build-directory output paths and the disjoint package policy
+that a real plugin configuration has to satisfy. The work is a source-set carve-out and a plugin
+block, not a redesign.
+
+**The fixtures use a text codec.** `GrpcTextCodec` gives the testkit a UTF-8 marshaller so the
+in-process and Netty lanes can exercise interceptors, status mapping, metadata limits and stream
+sequencing without generated stubs. Those contracts are properties of the platform and the transport,
+not of any message shape, so the substitution costs nothing — and the lanes run today rather than
+after codegen lands.
@@ -0,0 +1,50 @@
+# ADR-GRPC-003: Transport, business and stream evidence are three axes, and none implies another
+
+- Status: accepted
+- Date: 2026-08-30
+- Scope: `:grpc:grpc-core-api`, `:grpc:grpc-policy`, `:grpc:grpc-testkit`
+
+## Context
+
+A failed RPC produces a status code, and a status code is not an answer to the question the caller
+actually has. `DEADLINE_EXCEEDED` on a mutation does not say whether the mutation happened;
+`UNAVAILABLE` after the request was sent does not say the server never saw it; response headers
+arriving does not say a transaction committed.
+
+Every one of those is a place where a plausible inference produces a duplicate write or a lost one,
+and none of them is visible in a test that only exercises the happy path.
+
+## Decision
+
+Model what happened as three independent axes, and refuse the inferences between them.
+
+`GrpcTransportEvidence` records what the client observed on the wire, and distinguishes `NOT_SENT` —
+the client watched its own send fail — from `UNOBSERVED`, which is every other case where nothing is
+known. `GrpcBusinessEvidence` records what the application confirmed, with `COMMIT_UNKNOWN` as a real
+state rather than a placeholder. `GrpcStreamEvidence` is a sealed hierarchy whose non-empty cases all
+carry a position, because "partial" without a last sequence can be neither resumed nor reconciled.
+
+`GrpcExecutionEvidence` holds all three and rejects combinations nobody could have observed: a unary
+call with stream evidence, or a request the client watched fail to send that nonetheless carries
+business evidence. Promoting response headers to a confirmed commit is possible only by editing
+`withResponseHeadersSeen`, which is one method rather than a plausible line in an interceptor.
+
+`GrpcCompletionOutcome.forMutation` derives what a caller may conclude, and defaults
+`DEADLINE_EXCEEDED` and post-send `UNAVAILABLE` on a mutation to `COMPLETION_UNKNOWN`.
+
+The same types are used by the failure model and by the observation convention, so an incident has
+one account of a call rather than two.
+
+## Consequences
+
+**A whole class of retry bug becomes unrepresentable.** `GrpcRetryEligibility` reads all three axes
+plus the idempotency profile; a caller cannot reach "retry" from a status alone because the status
+alone is not an input.
+
+**The fault lane has something to check.** `GrpcTransportEvidenceClassifier` turns a client's
+observations into evidence and refuses to infer `NOT_SENT` from an unobserved state — and the lane
+exercises it against a real connection dropped mid-call, not against a mock.
+
+**Callers must handle a third outcome.** `COMPLETION_UNKNOWN` is not a failure and not a success, and
+a caller that treats it as either is wrong. `GrpcOperationStatusQuery` and `GrpcCompletionReconciler`
+exist so that resolving it is a supported path rather than an exercise for the caller.
@@ -0,0 +1,52 @@
+# ADR-GRPC-004: One retry owner, and keyed mutations need a durable ledger
+
+- Status: accepted
+- Date: 2026-08-30
+- Scope: `:grpc:grpc-policy`, `:grpc:grpc-operation-ledger-jpa`, `:grpc:grpc-core-api`
+
+## Context
+
+Three layers can retry a gRPC call: the application, the channel's service config, and a service
+mesh. Their effects multiply. Three attempts at each layer is twenty-seven requests for one call, and
+the load arrives exactly when the dependency is already failing.
+
+Separately, a mutation that is safe to repeat needs somewhere to record that it ran. Without one, a
+retry after a lost response either duplicates the effect or drops it, and nothing distinguishes the
+two afterwards.
+
+## Decision
+
+**Exactly one retry owner per channel.** `GrpcRetryOwner` has four values including `NONE`, which is a
+decision rather than an omission. `GrpcServiceConfigPolicy` refuses an in-process retry entry when the
+owner is the mesh or nobody, and `GrpcRetryOwnershipValidator` compares the service config's method
+names against the policy catalog — a renamed method leaves its retry entry matching nothing, silently,
+and the method then runs with channel defaults.
+
+**Retry eligibility reads the method, the evidence and the status together.**
+`GrpcRetryEligibility` refuses a non-idempotent method outright, refuses any call whose stream
+delivered a prefix, and turns a `DEADLINE_EXCEEDED` or post-send `UNAVAILABLE` mutation into
+"resolve the completion first" rather than a retry.
+
+**A keyed mutation is retryable only with both a caller key and a durable ledger.**
+`GrpcOperationLedger` is a port in `grpc-core-api`, so the policy layer can require durable
+idempotency without depending on a database. Its `claim` contract is a single atomic insert-or-read
+against a unique constraint: `JpaGrpcOperationLedger` inserts first and reads on constraint violation,
+because a read-then-insert implementation has a window exactly as wide as the race it closes and
+passes every test that does not run two attempts concurrently.
+
+The identity is caller fingerprint plus full method plus hashed key. All three are load-bearing:
+without the caller, one tenant's key suppresses another's write; without the method, a key reused
+across operations makes the second a replay of the first.
+
+## Consequences
+
+**A budget bounds retries as a fraction of traffic.** `GrpcRetryBudget` degrades to roughly no
+retries when everything is failing, which is the behaviour that lets a dependency recover.
+
+**The ledger and the mutation should commit together.** `JpaGrpcOperationLedger` carries no
+transaction annotations, deliberately: a `REQUIRES_NEW` would put the claim in its own transaction and
+reintroduce the window where the write is durable and the claim is not.
+
+**A key reused for a different request is a caller error, not a duplicate.** The stored request
+fingerprint turns that into `FAILED_PRECONDITION` rather than silently returning the first request's
+answer.
@@ -0,0 +1,50 @@
+# ADR-GRPC-005: One writer per stream, a bounded queue, and resume that refuses to guess
+
+- Status: accepted
+- Date: 2026-08-30
+- Scope: `:grpc:grpc-policy`
+
+## Context
+
+`StreamObserver` is not thread-safe, and the failure when two producers call `onNext` concurrently is
+not an exception — it is interleaved bytes, which a client decodes as a corrupt message or, worse, as
+a valid one it should never have received.
+
+Two further properties of server streams are easy to get wrong in ways that look healthy. A consumer
+that falls behind either terminates the stream or silently loses messages, and the second leaves a
+client with a stream that appears fine and is missing changes. And a reconnect either continues from
+a position the server can still replay, or skips whatever is no longer there.
+
+## Decision
+
+**A bounded queue drained by one writer.** `GrpcSerializedStreamWriter` accepts messages from any
+thread and hands them to the transport only from `flush`, which is synchronized. `write` returning
+`ACCEPTED` means queued, and the name is deliberately not `sent`: the transport call returns as soon
+as bytes are handed over, so no method here can honestly report delivery.
+
+**Both a message bound and a byte bound.** Either alone is unbounded in the other dimension.
+`GrpcFlowControlPolicy` also takes the transport's own readiness signal, because a writer that relies
+only on its queue bound produces as fast as it can allocate.
+
+**Termination is the default for a slow consumer.** `GrpcSlowConsumerPolicy.DROP_OLDEST` exists for
+feeds whose business meaning tolerates loss, and is not the default, because a client cannot detect
+dropped messages: the sequence numbers it sees are the ones it was sent.
+
+**Resume is refused rather than faked.** `GrpcStreamGapDetector` requires a signed, unexpired token
+whose caller and filter fingerprints match the current request, refuses one whose snapshot version
+moved, and returns `FULL_RESYNC_REQUIRED` when the cursor predates retained history. `GrpcResumeToken`
+carries a key id so the signing key can rotate without invalidating every outstanding token.
+
+## Consequences
+
+**A stream carries an envelope, not a bare payload.** `GrpcStreamEnvelope` holds the stream id,
+generation, sequence, snapshot version and resume token, because resume, gap detection and drain all
+need a position and a generation.
+
+**Four clocks, not one.** `GrpcStreamLifetimePolicy` separates setup deadline, idle timeout, max
+duration and heartbeat interval, and refuses combinations where one can never fire. Merging any pair
+produces a familiar bug: an idle timeout used as a max duration kills healthy busy streams.
+
+**A heartbeat is a liveness signal and nothing else.** It is not an application acknowledgement and
+not an ordering guarantee; `GrpcStreamHeartbeat` says so in the place somebody would otherwise reuse
+it.
@@ -0,0 +1,68 @@
+# ADR-GRPC-006: Stable discovery is DNS and static, and a Kubernetes profile names who balances
+
+- Status: accepted
+- Date: 2026-08-31
+- Scope: `:grpc:grpc-discovery`, `:grpc:grpc-client`
+
+## Context
+
+A gRPC channel's discovery configuration has a failure mode with no runtime symptom: it works, and
+it does not do what the dashboard says it does.
+
+The specific case is `round_robin` over a Kubernetes Service ClusterIP. The Service is one virtual
+address, so the resolver returns one endpoint and the client-side balancer has nothing to rotate
+across; kube-proxy picks a pod at connect time, and an HTTP/2 connection is long-lived, so every
+request from that client goes to the same pod for the life of the connection. Nothing fails. The
+configuration says `round_robin`, the metrics show requests spread across clients rather than pods,
+and the conclusion "we have client-side load balancing" is wrong in a way nobody is prompted to
+check.
+
+The mirror-image mistake is `pick_first` over a headless record, which pins a client to one pod out
+of many.
+
+Separately, a service mesh changes who owns retries, and a deployment that adds mesh routing without
+removing its own retry policy has two retriers whose effects multiply.
+
+## Decision
+
+**Stable resolvers are Static, DNS and Unix domain socket; Stable load balancing is `pick_first` and
+`round_robin`.** `GrpcDiscoveryPolicyValidator.requireStableScheme` refuses `xds`, `consul`, `etcd`
+and `eureka` by name, with a message saying they are Advanced capabilities with their own control
+plane and promotion gate rather than unknown schemes.
+
+**The pairing is checked against the resolved address count, not against intent.**
+`GrpcResolverProfile` carries `expectedAddressCount`, and `GrpcStableLoadBalancer.effective` answers
+whether the policy distributes anything over that many endpoints. A `round_robin` profile over one
+address is a reported violation whose message says it describes spreading that is not happening.
+
+**A Kubernetes deployment names its routing mode**, and the mode implies both the balancer and the
+retry owner. `GrpcKubernetesRoutingMode` has three values — `K8S_VIP`, `K8S_HEADLESS`, `MESH` — and
+`GrpcKubernetesProfile` refuses a mesh profile whose retry owner retries in-process.
+
+**A profile that carries long-lived streams must state a reconnect budget and a readiness drain
+grace.** A stream pins a client to one pod for its whole life, so every rollout, eviction and
+scale-down ends it. `GrpcKubernetesProfileValidator` additionally reports a VIP profile carrying
+long streams, and a drain grace shorter than the reconnect budget — the second means the pod stops
+serving before its clients have finished reconnecting elsewhere.
+
+**A DNS profile must refresh.** `GrpcResolverProfile` refuses a zero refresh interval on DNS,
+because a channel that resolved once at startup keeps sending to addresses that stopped existing an
+hour ago, and the resulting `UNAVAILABLE` looks like an unhealthy deployment long after the rollout
+finished.
+
+## Consequences
+
+**Two validators, not one.** `GrpcDiscoveryPolicyValidator` asks whether a balancer does anything
+over the addresses it will see; `GrpcKubernetesProfileValidator` asks whether the deployment shape,
+the retry owner and the stream obligations agree. A deployment can have a coherent resolver profile
+and still have put retries in two places, so merging them would let one answer hide the other.
+
+**`expectedAddressCount` has to come from somewhere.** It is a declared number, and a declaration can
+be wrong. It is still better than the alternative, which is not comparing anything: a wrong
+declaration is a wrong statement somebody wrote down, and a missing one is a question nobody asked.
+`GrpcChannelProfileValidator` takes resolved counts where they are known at startup and skips the
+check where they are not, rather than guessing and failing on a name that cannot be resolved yet.
+
+**xDS is reachable, and not by this route.** It lives in `grpc-advanced-resilience` behind its
+capability flag and its production approval, and `GrpcXdsStartupGuard.advertisableAsStableSupport()`
+returns false so the Stable support statement cannot widen quietly. See ADR-GRPC-ADV-001.
@@ -0,0 +1,55 @@
+# ADR-GRPC-ADV-001: Each advanced capability has its own flag, its own grade and its own promotion
+
+- Status: accepted
+- Date: 2026-08-30
+- Scope: `:grpc-advanced:*`
+
+## Context
+
+The advanced plan covers sixteen capabilities that differ by orders of magnitude in what they bring
+with them. gRPC-Web adds a proxy. Reactor adds a dependency. xDS adds a control plane, its outage
+modes, its own security boundary and its own version skew. Hedging duplicates production traffic.
+
+Bundling them under one flag makes enabling the cheapest of those the same decision as enabling the
+most consequential.
+
+## Decision
+
+**One flag per capability**, under `ca-skeleton.grpc.advanced.<capability>.enabled`, all off by
+default.
+
+**Four grades.** `ADVANCED_STABLE` starts on its flag; `EXPERIMENTAL` additionally needs a separate
+production approval, because the flag says somebody wanted the feature and the approval says somebody
+accepted that its failure modes are not fully characterised; `WATCH` cannot start at all; `DISABLED`
+is withdrawn.
+
+`GrpcAdvancedModuleGuard` distinguishes the three refusals — flag unset, grade unstartable,
+production unapproved — because the remedy differs in each case.
+
+**Promotion evidence is per capability.** `GrpcAdvancedPromotionEvidence` is one record per
+capability, so no promotion can drag another along;
+`GrpcAdvancedPromotionGate.capabilitiesDraggedAlong` returns an empty list, and that is a tested
+property rather than a claim. Two thresholds: seven days of soak plus complete evidence for
+`ADVANCED_STABLE`, thirty for a Stable default, because the second means every deployment gets the
+capability's dependencies and its failure modes.
+
+**Infrastructure is named per capability.** `GrpcAdvancedInfrastructureTestkit` records that
+gRPC-Web needs a proxy, Servlet needs a container, xDS needs a stoppable control plane and Kotlin
+needs a toolchain. A suite that runs without its infrastructure passes and establishes nothing, which
+is worse than not having one.
+
+## Consequences
+
+**The Kotlin adapter fails closed here, and says why.** This repository has no Kotlin toolchain, so
+`GrpcKotlinCompatibilityGate.supportableHere()` returns false. The four contract requirements — one
+schema source, coroutine cancellation propagation, Flow backpressure inside the Stable bounds,
+platform evidence types preserved — are checkable and are checked; only the compile lane is missing.
+
+**Edition 2026 cannot be used however its watch report reads.** `GrpcEdition2026Guard` is not
+conditional on the report, because letting a status record also authorise use means a schema moves
+onto an edition the moment somebody marks four fields SUPPORTED, with no promotion decision, no
+consumer migration and no ADR.
+
+**xDS is not part of the Stable support statement.** It works, behind its flag and its approval;
+`GrpcXdsStartupGuard.advertisableAsStableSupport()` returns false so a support matrix cannot widen
+quietly.
@@ -0,0 +1,35 @@
+# ADR-JPA-001 — The domain owns the persistence model
+
+- Status: Accepted
+- Date: 2026-08-11
+- Design: §10.1, §23.3
+
+## Context
+
+A persistence platform can either own the repository abstraction — a `GenericRepository<T, ID>`
+every aggregate inherits — or provide only the pieces domains assemble themselves.
+
+## Decision
+
+The domain owns entities, embeddables, repositories, queries, index requirements, and lock,
+soft-delete, and audit policy. The platform provides no generic CRUD repository and no base
+repository. `JpaRepositoryFragmentSupport` exists, has no `save`, `findById`, `findAll`, or
+`delete`, and is enforced not to acquire them.
+
+## Consequences
+
+A generic base repository has one property that looks like a benefit and is not: every aggregate
+gets the same operations. That means each aggregate is offered operations that may be wrong for it —
+a `delete` on an append-only ledger, a `findAll` on a table that will never be small — and, worse,
+one aggregate's later requirement changes the shared base and therefore changes behaviour for
+aggregates nobody reviewed.
+
+Spring Data already implements CRUD. Re-implementing it adds a layer whose only function is to be
+harder to opt out of.
+
+The cost is a small amount of repetition: each domain declares the repository interface it needs.
+That repetition is the thing that makes each aggregate's persistence surface reviewable.
+
+## Enforcement
+
+`JpaArchitectureRules.noGenericRepository()`; `JpaRepositoryFragmentSupportTest`.
@@ -0,0 +1,37 @@
+# ADR-JPA-002 — Retry re-runs the whole use case
+
+- Status: Accepted
+- Date: 2026-08-11
+- Design: §19.2
+
+## Context
+
+Optimistic conflicts, deadlocks, and serialization failures are recoverable. The question is what
+unit gets retried: the failed statement, the transaction, or the use case.
+
+## Decision
+
+The whole use case, in a new transaction with a new Persistence Context.
+`FullTransactionRetryCoordinator` re-enters `JpaTransactionExecutor` for every attempt, and the
+retry advice is ordered outside Spring's transaction advice so each attempt begins a new
+transaction.
+
+## Consequences
+
+Statement-level retry is wrong for exactly the failures being retried. An optimistic conflict means
+the state the attempt computed against is no longer the committed state; re-issuing the same
+statement computes the same wrong answer against a version that has moved on. The domain rules have
+to run again over reloaded data, which means the whole use case.
+
+Reusing the Persistence Context would be equally wrong: the second attempt would read the first
+attempt's stale entities out of the first-level cache. And with the advice ordering inverted, the
+retry loop would run inside one transaction that has already been marked rollback-only, so the
+second attempt fails immediately without executing anything.
+
+The cost is that a retryable use case must be safe to run from scratch — no irreversible external
+effect before the commit. `IrreversibleSideEffectContext` lets a use case declare when that does not
+hold, and the policy then refuses to retry it whatever budget remains.
+
+## Enforcement
+
+`FullTransactionRetryCoordinatorTest`; `RetryableJpaTransactionInterceptor.DEFAULT_ORDER`.
@@ -0,0 +1,41 @@
+# ADR-JPA-003 — Completion unknown is never retried
+
+- Status: Accepted
+- Date: 2026-08-11
+- Design: §17
+
+## Context
+
+A connection can break while a commit is in flight. The server may have committed; the
+acknowledgement may simply have been lost. The driver cannot tell the two apart.
+
+## Decision
+
+`TransactionCompletionUnknownException` is never retried, automatically or otherwise. It is
+produced only by a failure observed while the transaction phase is `COMMITTING`, and only for
+SQLSTATE `40003`, a connection-class (`08*`) state, or a transport break. Recovery is
+domain-specific reconciliation through `TransactionCompletionResolver`.
+
+## Consequences
+
+Retrying a possibly-committed write is the most damaging thing this platform could do: a duplicate
+payment, a duplicate order, a double decrement. There is no budget or backoff that makes it safe,
+because the failure is epistemic rather than transient.
+
+The invariant is enforced at the type level rather than by policy alone. `JpaFailureContext` refuses
+to construct a retryable completion-unknown context, and the exception rebuilds its context through
+the safe factory whatever it is handed. A future policy bug therefore cannot produce an unsafe
+retry — the value it would need does not exist.
+
+The rule is deliberately narrow in the other direction too. Classifying every connection failure as
+completion-unknown would push ordinary pool exhaustion and server restarts into the reconciliation
+queue, which trains operators to clear that queue without reading it — and then the one entry that
+mattered gets cleared with the rest.
+
+The cost is that the domain must supply the resolver. The platform cannot: only the domain knows
+which idempotency record, business row, or outbox entry proves the write happened.
+
+## Enforcement
+
+`JpaFailureContextTest`; `DefaultJpaRetryPolicyTest`; `CommitFailureClassifierTest`; release gate
+`completion-unknown-no-retry`.
@@ -0,0 +1,37 @@
+# ADR-JPA-004 — Flyway is the schema source of truth
+
+- Status: Accepted
+- Date: 2026-08-11
+- Design: §31
+
+## Context
+
+Hibernate can create and alter schema from the entity mapping. Flyway can apply versioned scripts.
+Both cannot own the schema.
+
+## Decision
+
+Flyway owns every schema change. Hibernate validates and never mutates: `ddl-auto` is `validate` or
+`none`, enforced at startup. The runtime database credential holds no DDL privilege, so the rule is
+enforced by the server as well as by configuration.
+
+## Consequences
+
+`ddl-auto=update` fails in a specific and expensive way: it adds but never drops or narrows, so the
+result is a schema that is neither the previous one nor the one the mappings describe — produced
+silently, by whichever instance started first, with no record of what it did.
+
+Two credentials rather than one is what makes this more than a convention. A configuration rule can
+be overridden by a property; a role without `CREATE` cannot be overridden by anything the
+application does.
+
+Validation fails closed and never repairs. `repair` rewrites the schema history to match the scripts
+on disk, which resolves a checksum mismatch by deleting the evidence of which change is missing.
+
+The cost is that a schema change requires a migration script and a deployment step. That is the
+intended cost: it makes schema change reviewable and reversible.
+
+## Enforcement
+
+`JpaDangerousConfigurationGuard`; `FlywaySchemaPolicy`; `FlywayValidationGate`;
+`PostgreSqlRuntimeRoleVerifier`; release gates `flyway-validate` and `runtime-role-no-ddl`.
@@ -0,0 +1,38 @@
+# ADR-JPA-005 — Contracts run against real PostgreSQL
+
+- Status: Accepted
+- Date: 2026-08-11
+- Design: §40
+
+## Context
+
+An in-memory database makes tests fast and hermetic. A container makes them slow and requires
+Docker.
+
+## Decision
+
+Every persistence contract runs against real PostgreSQL 16, 17, and 18 in containers. H2 remains a
+local-development convenience and never satisfies a contract. The lanes fail closed when Docker is
+absent rather than skipping.
+
+## Consequences
+
+The behaviours these contracts verify either do not exist in H2 or differ there: SQLSTATE values for
+the same violation, `FOR UPDATE SKIP LOCKED` semantics, JSONB operators, range types, concurrent
+index builds, `search_path` privileges, and the generated SQL for a paged collection fetch. A green
+H2 run is evidence that the code compiles and runs — not that any of the above holds.
+
+Three versions rather than one because the platform claims three. A contract suite that ran only on
+16 would make "Stable on 17 and 18" an assumption.
+
+Skipping on missing Docker is the failure mode this decision most wants to avoid: a skipped contract
+reports success, and CI eventually inherits that silence. `PostgreSqlContainerFactory.assertDockerAvailable()`
+throws instead.
+
+The cost is that the contract lanes need Docker and take minutes. The unit lane stays hermetic and
+fast, and is where most tests live; the container lanes verify the things only a real server can
+answer.
+
+## Enforcement
+
+`PostgreSqlVersion.stable()`; `PostgreSqlContainerFactory`; release gate `postgresql-contract`.
@@ -0,0 +1,67 @@
+# ADR-JPA-006 — `audit` is the canonical technical audit model; `auditing` stays a frozen candidate
+
+- Status: Accepted
+- Date: 2026-08-24
+- Review: `docs/reviews/2026-08-14-jpa-module-code-review.md` JPA-022
+
+## Context
+
+Two complete technical-audit mechanisms live in this leaf and they disagree about the schema.
+
+`audit/AuditableEntity` stamps `created_at`/`created_by`/`updated_at`/`updated_by` with an actor
+column of length 256, captured through explicit `initializeAudit`/`applyModification` calls and an
+`AuditContextPort`. `auditing/AuditMetadata` is a Spring Data embeddable that stamps
+`created_*`/`modified_*` with an actor column of length 64, captured by `@CreatedDate` and friends
+through an `AuditorAware`.
+
+Only the first is real: it is what the sample entities extend and what the migrations were written
+for. `JpaAuditingConfiguration` is not a Spring `@Configuration`, and nothing in production
+constructs any of the three `auditing` types.
+
+The review asked for one canonical model with a migration or activation decision. The failure mode
+it was protecting against is specific: an author of a new entity picks whichever package they find
+first, and column names, actor lengths and capture lifecycles then diverge per table.
+
+## Decision
+
+`audit/AuditableEntity` is canonical. `auditing` stays in the tree as a candidate and is excluded
+from the Stable capability report.
+
+The candidate is not deleted and not promoted. Deleting it would discard a working Spring Data
+integration that a deployment preferring declarative auditing would want. Promoting it would mean
+either renaming `modified_*` to `updated_*` and widening the actor column — a schema migration of
+every audited table to gain nothing a caller asked for — or moving the sample entities onto
+`modified_*`, which is the same migration in the other direction.
+
+Neither is worth doing now. What the divergence actually needed was not consolidation but a rule
+that an entity cannot straddle the two, and that rule is cheaper than either migration.
+
+## Consequences
+
+Two audit mechanisms remain readable in one leaf, and a reader has to be told which one is live.
+That cost is paid in this document, in the package javadoc and in a test whose name says so.
+
+Two failure modes stay silent unless they are asserted, so both are:
+
+- The candidate acquires a stereotype and starts stamping in every deployment that has this module
+  on the classpath, including the ones whose tables have no `modified_*` columns — where the result
+  is a failed startup rather than a feature.
+- Somebody "harmonises" the two by editing one side's column names, at which point the schema a
+  deployed table was migrated for and the schema its entity expects diverge with no migration
+  between them.
+
+If the candidate is ever promoted, it is promoted atomically: forward migration, sample conversion,
+`AuditContextPort → AuditorAware` and `Clock → DateTimeProvider` bridges land together, and this
+ADR is superseded rather than amended.
+
+Bulk and native updates stamp nothing under either mechanism. That is a property of JPA, not of the
+choice made here, so it is enforced separately rather than assumed away.
+
+## Enforcement
+
+`JpaAuditMechanismRule.entitiesUseExactlyOneAuditMechanism` and
+`bulkUpdatesOfAuditedEntitiesStampAudit`, run against the real production graph by
+`JpaProductionArchitectureTest` at the composition root — not against fixtures, which is how the
+earlier version of this rule pack passed while applying to nothing. `AuditingCandidateStatusTest`
+asserts the candidate carries no composing stereotype and that the two column sets stay distinct.
+`JpaAuditMechanismRuleTest` exercises the rules' own negative cases.
@@ -0,0 +1,63 @@
+# ADR-MONGO-001 — MongoDB platform boundary
+
+- **Status:** Accepted
+- **Date:** 2026-08-13
+- **Design source:** `docs/superpowers/specs/2026-08-11-mongodb-document-persistence-platform-design.md` §1, §2 (D-01, D-04, D-05), §5, §6
+
+## Context
+
+Two failure modes are common when a team wraps MongoDB.
+
+The first is flattening: a shared `CommonMongoRepository<T, ID>` and a generic CRUD facade, which
+forces every collection to share an id strategy, a consistency profile and a query surface. MongoDB's
+single-document atomicity, aggregation model and change streams stop being reachable, and the first
+collection that needs something different gets a cast or a leaky generic.
+
+The second is unrestricted exposure: the driver and `runCommand` available everywhere. Then any
+service can drop a collection, run an unbounded pipeline, or issue an admin command from a request
+thread, and no review catches it because there is nothing structural to catch.
+
+## Decision
+
+The domain owns its documents; the platform owns the cross-cutting decisions. Four exposure planes:
+
+| Plane | Contents | Client |
+|---|---|---|
+| D1 Standard document persistence | Spring Data repositories, typed queries, mapping manifest, atomic update primitives, optimistic revision | Stable API V1, `apiStrict=true` |
+| D2 Advanced document operations | `MongoTemplate`, transactions/sessions, bulk, aggregation, keyset cursors, change streams | Stable API V1, `apiStrict=true` |
+| D3 Explicit Mongo capability | Native BSON, time series, search/vector, CSFLE/QE, shard-aware operations | Separate capability client |
+| D4 Admin plane | Collection, validator, index, migration, shard, repair | Separate admin client and credential |
+
+Specifically:
+
+1. **No `CommonMongoRepository<T, ID>`.** Each aggregate declares its own repository.
+2. **D1/D2 run on Stable API V1 with `apiStrict=true`,** so a command outside the versioned API fails
+   at development time instead of on the next server upgrade.
+3. **D3 is not a raw-client escape.** Every call passes a fixed admission order: capability registered
+   → database profile → collection allowlist → operation name → timeout → consistency profile →
+   result limit → trace → redaction → command category → admin-command refusal → execute.
+4. **D4 is a separate client with a separate credential.** No application-plane path reaches it;
+   `PolicyAwareMongoNativeGateway` refuses admin-category commands regardless of capability.
+5. **Advanced and Experimental capabilities are opt-in modules**, never transitive dependencies of the
+   Stable surface.
+
+## Consequences
+
+**Positive.** MongoDB's semantics stay reachable. Misuse is refused structurally rather than reviewed
+for. A server upgrade cannot silently change D1/D2 behaviour. Admin operations have their own audit
+trail and credential.
+
+**Negative.** Every operation needs a registered name and profile, so a new query is a small amount of
+configuration rather than zero. A genuinely new capability requires a registration before it can be
+used. Both are deliberate: the cost is paid once per operation, at review time.
+
+**Rejected alternative — "expose the driver, rely on code review."** Review does not scale to every
+query in every service, and the operations that matter (unbounded pipeline, `dropCollection`,
+unanchored regex on user input) look unremarkable in a diff.
+
+## Repository adaptation
+
+The design assumes 19 Gradle modules under `modules/mongodb/`. This repository's fail-closed registry
+declares exactly 19 leaf identities, so the modules became package boundaries inside
+`:adapter:outbound:persistence-mongo`, enforced by ArchUnit. See
+[docs/mongodb/repository-adaptation.md](../mongodb/repository-adaptation.md).
@@ -0,0 +1,58 @@
+# ADR-MONGO-002 — BSON representation is a pinned manifest
+
+- **Status:** Accepted
+- **Date:** 2026-08-13
+- **Design source:** design §10, decision D-06
+
+## Context
+
+How a Java value is represented in BSON is a data contract, but nothing in the default toolchain
+treats it as one. Spring Data and the MongoDB driver both have defaults, and those defaults have
+changed across versions. A `BigDecimal` can land as a `Double`, a `String` or a `Decimal128`; a `UUID`
+can land as `Binary` subtype 3 or subtype 4; an `Instant` can land as a `Date` or a `String`.
+
+The consequences are asymmetric. A representation change is invisible in a value-equality test —
+`12.30` looks like `12.30` whether it is a double or a `Decimal128` — but once a collection holds
+production data, changing it is a full migration. And the UUID case is worse than a migration: legacy
+Java representation byte-swaps two halves of the UUID, so a document written under one representation
+and read under the other yields a *different, valid-looking* UUID. Nothing errors. You get the wrong
+record.
+
+## Decision
+
+`MongoTypeRepresentationManifest` pins the representation for every type the platform maps, and
+`MongoMappingConfiguration` builds the Spring Data converters from it. Nothing relies on a library
+default.
+
+| Java | BSON | Rationale |
+|---|---|---|
+| `UUID` | `Binary` subtype 4 (`STANDARD`) | Subtype 3 byte-swaps; cross-representation reads are silently wrong. |
+| `BigDecimal` | `Decimal128` | A double cannot represent `12.30`; money compared as a double is eventually wrong by a cent. |
+| `BigInteger` | `Decimal128`, or declared `String` when out of range | 34 significant digits; out of range fails on write instead of rounding. |
+| `Instant` / `OffsetDateTime` / `ZonedDateTime` | UTC `Date` | One instant, one representation. |
+| `LocalDate` | declared per field | A calendar day is not an instant. |
+| `LocalDateTime` | **refused** | No offset: the stored value depends on the writing JVM's default zone. |
+| `enum` | `String` name | Ordinals renumber when someone inserts a constant. |
+
+Type metadata follows `MongoTypeMetadataPolicy` — `NONE`, `ALIAS` or `CLASS_NAME`. A
+`@LongLivedMongoDocument` type may not use `CLASS_NAME`: writing a FQCN into a million documents makes
+a package rename a data migration.
+
+The manifest is enforced by a golden gate. `MongoBsonSnapshot` canonicalises a stored document,
+preserving BSON types and keeping missing distinct from null, and
+`MongoBsonSnapshotAssert.hasTypeSignature(...)` fails on any representation change. The registry
+pins `UuidCodec(STANDARD)` explicitly rather than inheriting a default, since inheriting the default
+is the exact drift the gate exists to catch.
+
+## Consequences
+
+**Positive.** A library upgrade cannot move a representation without failing a test. Money is exact.
+UUIDs read back as themselves. Class moves stay refactors.
+
+**Negative.** Every representation-affecting change requires updating a snapshot *and* writing a
+migration. A new mapped type needs a manifest entry before it can be used. This is the intended
+friction: the alternative is discovering the change in production.
+
+**Rejected alternative — "snapshot the JSON."** JSON destroys exactly the distinctions the gate
+protects: `Decimal128` and `String` both render as text, `Binary` UUID and `ObjectId` both render as
+hex, and missing and null both disappear.
@@ -0,0 +1,68 @@
+# ADR-MONGO-003 — Transaction body retry and commit retry are separate loops
+
+- **Status:** Accepted
+- **Date:** 2026-08-13
+- **Design source:** design §14–§16, decisions D-07 through D-10
+
+## Context
+
+MongoDB reports two transaction failures that look similar and must be handled in opposite ways.
+
+`TransientTransactionError` means the transaction definitively did not commit. The correct response is
+to run the whole thing again.
+
+`UnknownTransactionCommitResult` means the commit **may already have applied** — typically because the
+primary changed while the commit was in flight. The correct response is to retry *the commit*, which
+is a no-op if it already succeeded.
+
+The common implementation wraps everything in one retry loop. That loop replays the body after an
+unknown commit, and if the commit did apply, the body applies twice. In a payment or notification path
+that is a duplicate charge or a duplicate message, produced by the error handler.
+
+The related trap is session reuse: retrying on the same session after an abort carries the aborted
+transaction's state into the retry.
+
+## Decision
+
+`MongoTransactionRetryCoordinator` implements two loops with different scopes.
+
+```
+for each body attempt within the budget:
+    open a NEW session
+    run the body
+        TransientTransactionError -> abort, continue to next body attempt
+    commitWithRetry(session):
+        UnknownTransactionCommitResult -> retry the COMMIT ONLY, same session
+```
+
+Rules that follow, all of them load-bearing:
+
+1. **A new session per body attempt.** No aborted state leaks into a retry.
+2. **The body is never replayed after a commit ambiguity.** `MongoRetryScope.COMMIT_ONLY` is a
+   distinct value from `BODY` precisely so this cannot be collapsed by accident.
+3. **One budget bounds both loops.** `MongoRetryBudget` limits attempts *and* elapsed time, with
+   jittered backoff, so a struggling primary is not retried into the ground by every instance at once.
+4. **An exhausted commit retry surfaces `TRANSACTION_COMMIT_UNKNOWN`,** never a generic failure. An
+   ambiguous outcome reported as a failure invites the caller to retry — the one thing that must not
+   happen. See [docs/mongodb/runbooks/unknown-commit.md](../mongodb/runbooks/unknown-commit.md).
+5. **Transaction bodies write a deterministic marker** so `MongoCommitReconciler` can establish what
+   actually happened. A transaction that cannot be reconciled has no recovery path.
+6. **Classification reads labels before codes.** Server error labels are the authoritative statement
+   about retryability; error codes vary by version.
+
+Surrounding decisions that reduce how often this path is reached at all: single-document atomic
+operations are preferred over transactions (D-09), partial changes use update operators rather than
+`save()` (D-07), and whole-document replacement requires an optimistic revision (D-08).
+
+## Consequences
+
+**Positive.** A commit ambiguity cannot become a duplicate effect. The ambiguity reaches the caller as
+an ambiguity. The retry budget is bounded in both attempts and time.
+
+**Negative.** Callers must handle a third outcome beyond success and failure. Transaction bodies must
+write a marker they would not otherwise need. Both costs are small compared with reconciling
+duplicated financial effects after the fact.
+
+**Rejected alternative — "one retry loop, at-least-once everywhere."** It requires every transaction
+body to be fully idempotent, which is a much stronger and much less checkable property than writing
+one marker, and it is silently violated the first time someone adds a non-idempotent step.
@@ -0,0 +1,62 @@
+# ADR-MONGO-004 — Index and schema changes belong to the admin plane
+
+- **Status:** Accepted
+- **Date:** 2026-08-13
+- **Design source:** design §21–§25, decisions D-11, D-13
+
+## Context
+
+Spring Data can create indexes automatically from annotations. On a laptop this is convenient. On a
+collection with a hundred million documents, an index build is a capacity event: it consumes CPU, IO
+and memory on the primary for minutes to hours, and it starts because a pod restarted.
+
+Worse, it starts N times when N pods restart, and there is no approval step, no ordering relative to
+the code that needs the index, and no record afterwards of what was created.
+
+Schema validators have the same shape with a sharper edge: tightening a validator on a collection with
+existing data rejects writes to documents that were legal when they were written.
+
+TTL has a third shape. It looks like a scheduler and is not one: the TTL monitor runs about once a
+minute and deletes in batches, so an expired document routinely remains readable for minutes or hours.
+
+## Decision
+
+**Indexes and validators are declared in a manifest and applied by the admin plane (D4).** Automatic
+index creation in production is disabled.
+
+1. `MongoManifestRegistry` holds the declared indexes (`MongoIndexManifest`) and validator
+   (`MongoSchemaManifest`) per collection. The manifest is the source of truth, reviewed in a pull
+   request.
+2. `MongoIndexDiffEngine` compares manifest against observed state and reports missing, extra and
+   *changed* indexes. Changed ones are reported rather than re-issued: MongoDB will not silently
+   rebuild an index whose definition moved.
+3. `MongoIndexApplyPolicy` sets what an environment may do — `APPLY` (local), `APPLY_WITH_DIFF`
+   (staging), `DIFF_WITH_APPROVED_APPLY` (production), `REPORT_ONLY` (audit).
+4. **Ownership gates every drop.** `MongoMetadataOwnership` distinguishes `APPLICATION_MANAGED` from
+   `SEARCH_MANAGED`, `ENCRYPTION_MANAGED` and `EXTERNAL`. Only application-managed objects are
+   droppable on drift. A diff engine without ownership eventually proposes dropping
+   `enxcol_.customers.esc`, and "the drift tool cleaned it up" is a very bad incident summary.
+5. **Retirement is staged.** `MongoIndexRetirementState` moves an index declared → hidden →
+   observed-unused → droppable, one deployment per transition. Hiding is instantly reversible;
+   dropping is a rebuild.
+6. **Stable validation actions are `error` and `warn` only.** `errorAndLog` is not part of the Stable
+   contract on 7.0 or 8.0 and is refused. Tightening goes `warn`+`MODERATE` → confirm zero warnings →
+   `error`+`STRICT`, in two deployments.
+7. **TTL is physical cleanup only** (D-13). `MongoExpirationAccessPolicy` states the rule: a
+   document's presence is not authorization and its absence is not a deadline. Access control checks
+   the expiry field; scheduling uses a scheduler.
+8. **Migrations are checksummed, locked, precondition-checked and resumable.**
+   `MongoMigrationRunner` fails hard when an applied id's checksum changed — two environments running
+   different code under one id is worse than a failed deploy.
+
+## Consequences
+
+**Positive.** Index builds are scheduled by people who know the capacity. Rollback is possible at
+every step. Drift is visible without being dangerous. Nothing drops what it does not own.
+
+**Negative.** Adding an index is a manifest change plus an apply, not an annotation. Local development
+uses `APPLY` so the friction is confined to environments where it is warranted.
+
+**Rejected alternative — "auto-create with a feature flag."** The flag is either on in production,
+which is the problem, or off, in which case the manifest is the real mechanism and the annotation is a
+second, divergent source of truth.
@@ -0,0 +1,80 @@
+# ADR-MONGO-ADV-001 — Advanced capability promotion
+
+- **Status:** Accepted
+- **Date:** 2026-08-13
+- **Design source:** design §2 (D-15), §3.2–§3.3; Advanced expansion plan Task 15
+
+## Context
+
+Sharding, time series, CSFLE, Queryable Encryption, search, vector search and multi-tenancy each work
+in a demo within an afternoon. What they do not do is behave the same way in production, and the
+differences are not discovered by functional tests:
+
+- Sharding changes which queries are efficient. A query that misses the shard key becomes
+  scatter-gather, which passes every test on a one-shard cluster.
+- Encryption's failure modes are KMS failure modes — wrong key, revoked permission, mid-rotation —
+  none of which occur against a local key provider.
+- Search and vector search can be functionally correct and useless: the index returns results, and
+  the results are not relevant. Recall is not visible in a pass/fail assertion.
+- Database-per-tenant works until the tenant count crosses what the connection and file-handle
+  budget supports, which is an operational property, not a code property.
+
+The failure mode this ADR prevents is a capability marked "done" on the strength of a green test that
+never touched the environment where it will run.
+
+## Decision
+
+Every Advanced and Experimental capability is an **opt-in module behind its own flag**, and promotion
+requires evidence, not confidence.
+
+### Enablement
+
+`MongoAdvancedCapabilityFlags` gates construction of every Advanced entry point. A disabled capability
+does not produce a runtime warning — the type refuses to be constructed, naming the property that
+enables it (`MongoAdvancedCapabilityFlags.propertyFor(capability)`). Being on the classpath is not
+being enabled, and `stableNeverDependsOnAdvanced` (ArchUnit) keeps the Stable surface free of them.
+
+### Promotion evidence
+
+`MongoAdvancedPromotionGate.verify(evidence)` requires every category:
+
+| Category | Means |
+|---|---|
+| `stable-platform` | The Stable release gate passed on the same revision. |
+| `actual-topology` | The capability ran on the real topology — a real sharded cluster, the real KMS, the actual target deployment. Atlas Local is a pull-request convenience and explicitly not release evidence (`MongoAtlasCapabilityContractSuite.Environment.ATLAS_LOCAL`). |
+| `security` | Privileges reviewed; the capability's admin role is separate from the application role. |
+| `migration` | A documented path in and, where the capability is irreversible, an explicit statement that there is no path back. |
+| `failure` | Negative cases fail closed: wrong key, missing permission, rotation, non-ready index, unrouted query. |
+| `runbook` | A runbook exists for the capability's characteristic incident. |
+
+### Additional per-capability requirements
+
+- **Search / vector search:** relevance and performance evidence, not functional success alone.
+  `MongoVectorSearchBenchmarkGate` requires recall alongside latency and index size; a gate that
+  measures only latency certifies a fast wrong answer.
+- **Database-per-tenant and reshard orchestration remain Experimental** until operational scale
+  evidence exists. Both are correct in the small and unbounded in the large.
+- **Reshard requires an explicit `ReshardApproval`** — a named approver and a stated window. It
+  rewrites the collection.
+
+### Promotion does not change the dependency boundary
+
+A capability promoted to Stable **remains an opt-in module** unless a later starter ADR changes the
+dependency boundary. Promotion is a statement about evidence, not an invitation to add a transitive
+dependency to every service.
+
+## Consequences
+
+**Positive.** No capability reaches production on the strength of a container-only test. The evidence
+list is the same for every capability, so promotion is reviewable rather than negotiated.
+
+**Negative.** Promotion requires access to real infrastructure — a sharded cluster, a real KMS, the
+target deployment. That is the cost of the guarantee: the alternative is finding out in production,
+where encryption and sharding are both expensive to reverse.
+
+## Verification
+
+`scripts/verify-mongodb-advanced.sh` enforced this ADR until it was removed on 2026-08-15. The
+promotion evidence categories this ADR requires are therefore no longer checked by any automated
+gate; they are a review obligation until one is rebuilt. See `docs/mongodb/repository-adaptation.md`
+§5 for the Gradle lanes the script wrapped.
@@ -0,0 +1,69 @@
+# ADR-WEB-ADV-001: Streaming is live delivery, and the web module stores no history
+
+- Status: Accepted
+- Date: 2026-08-25
+- Scope: `adapter:inbound:web` — `advanced.stream.**`
+
+## Context
+
+Advanced Tasks 6–13 add SSE, NDJSON and JSON text sequences, with a `Last-Event-ID` resume path.
+
+One fact drives every decision here: **after the first byte, the HTTP status is 200 and cannot
+change.** A stream that ends because a dependency failed and one that ends because it finished are
+identical at the transport layer — both are a closed connection after a 200. So is a stream that was
+cut off mid-flight.
+
+The second fact is that a resume path invites the web module to remember things. It must not: the
+messaging platform already owns durable event history, and a second copy would have its own
+retention, its own eviction and its own opinion about ordering.
+
+## Decision
+
+**Three outcomes, expressed in the stream rather than in the status.** `WebStreamEnvelope` is sealed
+over `Item`, `Failure` and `Complete`. A client that sees neither terminal envelope has been cut off,
+and that third case is recorded as `ABRUPT_CLOSE` rather than counted as a completion — which is
+where a rising rate of mid-stream failures would otherwise hide.
+
+**Nothing writes a problem document onto a committed response.** `WebStreamTerminationMapper`
+branches on whether any byte has been written. Before commit, an RFC 9457 problem with a real
+status; after, a terminal record. Attempting both produces a body that is half stream and half JSON,
+which no client parses and every proxy caches as a success.
+
+**Positions are monotonic, and it is enforced.** `WebStreamEvidence.recordDelivered` refuses a
+repeated or regressing position. A client deduplicating on position would silently drop the second
+item.
+
+**A slow consumer is disconnected, not buffered.** `WebStreamPolicy.maxBufferedItems` is a hard
+bound. Backpressure protects the reactive pipeline; it does not protect the server's heap from a
+consumer that reads slowly for an hour.
+
+**Every stream is in a registry, and shutdown drains it.** A node with a hundred open streams and no
+other traffic looks idle by request rate. `WebStreamDrainCoordinator` stops accepting first, asks
+clients to reconnect, and only then forces the remainder — because a client whose socket is cut
+retries immediately, and if every socket is cut at once, every client retries at once.
+
+**The web module stores no durable history.** `WebStreamReplaySource` is an interface this module
+implements nowhere. An expired cursor raises `ReplayCursorExpiredException` rather than resuming from
+the oldest retained position, because that delivers a stream with a hole the client cannot see.
+
+**The replay-to-live seam is watched.** `GapAndDuplicateGuard` detects both directions. Neither is
+visible in either half on its own.
+
+## Consequences
+
+- Clients must handle three outcomes. A client that treats a closed connection as completion will be
+  wrong, and no server change can fix that for it.
+- An expired `Last-Event-ID` costs the client a full re-read. That is the honest answer.
+- JSON-seq is preferred over NDJSON where truncation matters: its separator comes first, so a parser
+  resynchronises at the next record. NDJSON's delimiter is the thing that gets truncated away.
+
+## Alternatives considered
+
+- **Emit a problem document when a stream fails after commit.** Rejected: the body becomes
+  unparseable and the 200 is cached.
+- **Resume from the oldest retained position when the cursor expires.** Rejected: positions are
+  contiguous from where the replay started, so nothing in the data says events are missing.
+- **Store replay history in the web module.** Rejected: a second source of truth that drifts
+  invisibly.
+- **Unbounded buffering for slow consumers.** Rejected: it moves the client's slowness into the
+  server's heap.
@@ -0,0 +1,66 @@
+# ADR-WEB-ADV-002: Virtual threads change scheduling, not the concurrency budget
+
+- Status: Accepted
+- Date: 2026-08-25
+- Scope: `adapter:inbound:web` — `advanced.virtualthread`, `advanced.blockingbridge`
+
+## Context
+
+Advanced Task 2 offers a virtual-thread executor for MVC; Task 3 offers a bounded blocking bridge
+for WebFlux.
+
+A platform-thread MVC deployment has an implicit concurrency limit — the thread pool — and that
+limit is usually what has been protecting the database pool, the outbound HTTP bulkhead and every
+downstream service from the full arrival rate. Nobody wrote it down as an admission policy; it was a
+side effect of the pool size.
+
+Switching to virtual threads deletes that limit without deleting anything that depended on it.
+
+## Decision
+
+**An explicit admission limit is required when virtual threads are enabled.**
+`VirtualThreadProfile` refuses construction without one. Without it the deployment accepts every
+arrival, queues all of them on the downstream budgets, and times out work that would have succeeded
+had it been refused. The load that used to be shed at the front door is shed at the back, after the
+cost of accepting it.
+
+**The limit bounds concurrent use cases, not threads.** `VirtualThreadAdmissionGuard` is a fair
+semaphore, not a pool. Bounding threads would put the waiting back and throw away what virtual
+threads bought. Ten thousand virtual threads may exist while a hundred hold permits.
+
+**The downstream budgets are carried in the profile and stated as unchanged.** The whole point is
+that they did not grow. `admissionFitsDownstreamBudgets()` reports when the admission limit exceeds
+them, without refusing — a deployment can legitimately admit more than its pool when the work is not
+all database-bound, and that should be a choice rather than an accident.
+
+**Blocking offloads are registered, bounded and timed out.** `boundedElastic()` is available from
+anywhere and unbounded in practice, so a controller that calls it has silently opted the whole
+application into an unbounded pool. `BlockingBridgeProfile` names the operations permitted to
+offload; `BlockingBridgeBudget` bounds the concurrency and refuses a caller that cannot get a slot
+in time, because otherwise a slow dependency's callers accumulate until the heap does and the fast
+dependencies starve behind them.
+
+**Pinning is observed, not assumed away.** `VirtualThreadProfile.requiredObservations()` lists what
+has to be watched — `jdk.VirtualThreadPinned` above all. A synchronized block held across a blocking
+call pins the carrier thread, the carrier pool is bounded by CPU count, and enough pinned carriers is
+a deadlock a thread dump does not obviously show.
+
+## Consequences
+
+- Enabling virtual threads is a two-part change: the executor and the admission limit. The profile
+  will not let it be one.
+- Refusals rise under load, and that is correct. A request refused in a millisecond is better for
+  the client than the same request accepted and timed out thirty seconds later behind a full pool.
+  An operator seeing 503s climb should read them as the limit working.
+- `VirtualThreadAdmissionGuard.peakActive()` exists so a load test can assert the limit was applied.
+  It is invisible from throughput, which is why a load test that only measures throughput would pass
+  with the guard removed.
+
+## Alternatives considered
+
+- **Enable virtual threads and raise the downstream budgets to match.** Rejected: the budgets are
+  sized to what the dependencies can serve, not to what the web tier can accept.
+- **Bound the virtual threads themselves with a fixed-size executor.** Rejected: that is a platform
+  thread pool with extra steps.
+- **Let controllers call `boundedElastic()` directly.** Rejected: every such call site is invisible
+  until the pool is the thing consuming the heap.
@@ -0,0 +1,55 @@
+# ADR-WEB-ADV-003: OpenAPI 3.2 is generated in parallel and stays experimental
+
+- Status: Accepted
+- Date: 2026-08-25
+- Scope: `adapter:inbound:web` — `advanced.openapi`
+
+## Context
+
+Advanced Task 17 adds an OpenAPI 3.2 generation lane beside the Stable 3.1.2 snapshot.
+
+Generating 3.2 is cheap. Adopting it is not, and the two get conflated because the generated
+document looks fine. The value of an API description is entirely in what consumes it, and a document
+in a version a client generator does not fully understand produces a client that compiles and is
+wrong — which is worse than no document at all.
+
+## Decision
+
+**3.1.2 remains the release artifact.** `OpenApiVersionLane.STABLE_3_1.releaseArtifact()` is true and
+`EXPERIMENTAL_3_2`'s is false. This is a property of the type, not a configuration setting.
+
+**Generating 3.2 must not change the 3.1 snapshot.** Both are produced from the same model, so a
+contributor that mutates it on the way to 3.2 changes the artifact that is actually shipped —
+silently, and only when the experimental lane runs. `OpenApi32CompatibilityReport` compares the
+snapshot hash before and after and makes a difference a promotion blocker.
+
+**Four kinds of tool are checked separately.** A parser reports structural errors; a linter applies
+style rules and accepts documents a parser rejects; a generator produces client code, and this is
+where an unsupported construct surfaces — not as an error but as a method with the wrong signature;
+a compile of that generated code is the only step that catches it. "OpenAPI 3.2 works" is not a
+statement anybody can make. "This document is read correctly by these four tools at these versions"
+is.
+
+**Promotion requires an accepted ADR regardless of how green the matrix is.**
+`promotionBlockers(false)` always contains that blocker. A machine-checkable matrix cannot decide
+whether the consumer population is ready.
+
+**Streaming description differences are reported separately.** They are the substantive difference
+between the two versions for this application, and folding them into a pass/fail hides what
+changed.
+
+## Consequences
+
+- The 3.2 document is published as an artifact of the experimental workflow, never of the release
+  workflow.
+- A client generator that only understands 3.1 is unaffected, which is the point.
+- Adopting 3.2 later is a documented decision with a named consumer matrix behind it.
+
+## Alternatives considered
+
+- **Switch to 3.2 and keep a 3.1 downgrade.** Rejected: the downgrade is lossy in exactly the
+  constructs 3.2 was wanted for, so it would ship a description that is wrong for both audiences.
+- **Generate only 3.2 and let consumers cope.** Rejected: the failure mode is a generated client
+  that compiles and misbehaves.
+- **Skip the client-compile step in the matrix.** Rejected: it is the only one that catches the
+  failure the others miss.
@@ -0,0 +1,56 @@
+# ADR-WS-001: The WebSocket platform ships as packages in one leaf, with machine-checked boundaries
+
+- Status: accepted
+- Date: 2026-08-25
+- Scope: `:adapter:inbound:websocket`
+
+## Context
+
+The realtime connection platform design models itself as eighteen Gradle modules under
+`modules/websocket`, each with a declared purity grade and a declared set of allowed dependencies.
+This repository's `src/config/architecture/modules.json` is a fail-closed registry that owns the
+leaf list; adding eighteen leaves is a registry change of a size that needs its own decision, and
+HARD-STOP #5 forbids doing it implicitly.
+
+Three earlier platforms in this repository — JPA, GraphQL, and the HTTP platform — met the same
+situation and resolved it the same way.
+
+## Decision
+
+The eighteen design modules ship as packages inside the single registered leaf. `WebSocketStableModule`
+declares each one's package, purity grade and exact allowed edges, and `WebSocketModuleBoundaryTest`
+scans the production tree and fails when the declaration and the tree disagree in either direction.
+
+Three deviations from the design's module map were forced by the check and are recorded in
+`docs/websocket/repository-adaptation.md`: `WebSocketSubprotocolName` moved to `core` and the codec
+moved to its own FRAMEWORK_BOUND module, both to avoid cycles the design's placement created here;
+and the `budget -> core` edge was inverted because `budget` imports nothing from `core`.
+
+## Consequences
+
+**The boundary is enforced, not documented.** Six violations were caught during implementation that
+a document would not have: two would-be cycles, a duplicate module declaration where two ids claimed
+one package, and three undeclared edges. The duplicate is the instructive one — with two ids on one
+package, ownership depends on iteration order and one module's rules silently apply to nothing. A
+guard against it is now part of the boundary test.
+
+**The detector had a hole.** Its framework-import list named `com.fasterxml` (Jackson 2) and not
+`tools.jackson` (Jackson 3), which is what Spring 7 actually uses — so a CORE module could have
+imported a mapper unnoticed. Fixed here and in the HTTP platform, which shared the list.
+
+**Promotion stays cheap.** Each enum constant is already shaped like a leaf specification, so
+splitting one out later is a registry edit rather than an archaeology exercise.
+
+**The design's own rules were kept where they cost something.** `core` names no framework, so the
+same decisions serve both runtimes and are testable without a server; no Java class name reaches the
+wire; the payload is an encoded string rather than a map; and handlers are given no way to write,
+which is what makes ordering and backpressure guarantees rather than conventions.
+
+## Alternatives considered
+
+**Register eighteen leaves.** Faithful to the design and a large change to a fail-closed registry
+for a platform that ships as one artifact either way. Rejected as disproportionate; the boundary
+test provides the property the modules were for.
+
+**Ship the modules as packages with no enforcement.** Cheapest, and it makes the boundary a claim.
+The six violations found during implementation are the argument against it.
@@ -0,0 +1,60 @@
+# ADR-WS-002: Resume and cluster state are caches, and are treated as caches
+
+- Status: Accepted
+- Date: 2026-08-25
+- Scope: `adapter:inbound:websocket` — `advanced.resume`, `advanced.cluster`, `advanced.presence`
+
+## Context
+
+Advanced Tasks 2–8 add three things that all look like state and are not: a resume token that says
+where a client got to, a cluster index that says which node holds a session, and a presence summary
+derived from that index.
+
+Each is a statement about the past. The resume token was minted before the disconnect; the index
+entry was written by a node that may since have died; presence is a read of the index and inherits
+everything wrong with it. The failure this ADR exists to prevent is treating any of them as current
+fact, because each reads as one at the call site.
+
+## Decision
+
+**Resume is bounded by what the replay store actually holds, not by what the token claims.**
+`ResumeCoordinator` consults `ReplayAvailability` before honouring a position. A token that names a
+position the store has evicted produces a resynchronise, not a gap-filled stream. The alternative —
+trusting the token — silently delivers a stream with a hole in it, which is worse than an explicit
+resynchronise because the client believes it is complete.
+
+**Cluster index entries carry an observation time and are checked against it on every read.**
+`ExternalSessionSummary.staleAt` exists so that "the index says edge-2" cannot be used without also
+answering "as of when". An entry whose node stopped reporting is not evidence that the node holds
+the session.
+
+**Durable fan-out is deduplicated by stream position, not by message id.** At-least-once is the
+contract, so redelivery is normal operation: a redeploy, a slow consumer or a broker rebalance all
+produce it. `FanoutDeduplicator` keys on `(stream, position)` and advances a high-water mark under
+`compute`, so two consumer threads cannot both deliver the same position.
+
+**Presence has four states, not two.** `OFFLINE` is a reported fact; `STALE` is the absence of one.
+Collapsing them reports every user as disconnected during a Redis partition, when what happened is
+that the index went dark and the connections are fine.
+
+**Nothing security-relevant may depend on presence.** An attacker who can make a node stop reporting
+can move the platform's belief about who is present. Presence answers "show a green dot".
+
+## Consequences
+
+- A resume that cannot be honoured is visible to the client as a resynchronise. Clients must
+  implement one; there is no mode in which the platform silently pretends.
+- Every read of the cluster index needs a clock. This is deliberate friction.
+- `PresenceSummary.classify` refuses an idle window at or past the stale window, because otherwise
+  `IDLE` is unreachable and the caller believes it has a four-state model when it has three.
+- Fan-out envelopes carry a bounded reference and the catalog-encoded document, never a business
+  object. A rolling deploy has two versions of the code reading the same envelope.
+
+## Alternatives considered
+
+- **Trust the resume token.** Rejected: it makes a gap indistinguishable from a complete stream.
+- **Deduplicate by message id.** Rejected: a broker that redelivers may re-mint ids, and a producer
+  that retries certainly does. Position is the property the ordering actually has.
+- **A single `online` boolean.** Rejected for the partition case above.
+- **Write presence separately from the session index.** Rejected: two sources of truth for "who is
+  connected" drift, and the drift is invisible — both look plausible and nothing reconciles them.
@@ -0,0 +1,74 @@
+# ADR-WS-003: STOMP is an Advanced adapter with a declared destination catalog
+
+- Status: Accepted
+- Date: 2026-08-25
+- Scope: `adapter:inbound:websocket` — `advanced.stomp`, `advanced.stomp.rabbit`
+
+## Context
+
+Advanced Tasks 9–13 add STOMP 1.2 alongside the platform's own protocol, plus a RabbitMQ broker
+relay and cross-node user destinations.
+
+This leaf already ships an older STOMP-over-SockJS channel (`stomp`, gated on
+`ca-skeleton.websocket.enabled`). Two `@EnableWebSocketMessageBroker` configurations in one context
+do not conflict loudly — both contribute a configurer, both call `configureMessageBroker`, and the
+broker that results is whichever ran last. Nothing errors and nothing logs.
+
+STOMP also brings a destination model that is a free string from the client. Without a catalog, the
+set of reachable destinations is whatever the broker accepts, which for the simple broker is every
+string.
+
+## Decision
+
+**The Advanced adapter is its own module (`advanced-stomp`), separate from `advanced`.** It is the
+one Advanced capability that cannot be pure — STOMP here *is* the Spring Messaging types — and
+folding it into `advanced` would relax that module's purity for every capability in it.
+
+**The relay is a further module (`advanced-stomp-rabbit`).** The adapter parses a protocol; the
+relay opens a TCP connection to somebody else's broker and makes every delivery depend on it.
+Different blast radius, different decision, different module.
+
+**Destinations are declared, per operation.** `StompDestinationCatalog` maps `(operation,
+destination)` to a required permission. Undeclared is refused. `SUBSCRIBE` and `SEND` are separate
+declarations, because reading a feed and publishing into it are different rights.
+
+**The authorization decision is a value, not an interceptor method.** `StompAuthorizationPolicy`
+returns a `StompAuthorizationDecision`; `StompSecurityInterceptor` only extracts and enforces. A rule
+reachable only through a `MessageChannel` gets tested for the cases somebody built a channel for.
+
+**Only one STOMP runtime may run.** `StompBrokerExclusivity` fails the context when both channels
+are enabled, when both brokers are, or when the adapter is enabled with no broker behind it.
+
+**A `RECEIPT` is never promoted to a commit.** `StompEvidence` has six stages and
+`StompAckPolicy.evidenceForReceipt()` is fixed at `PROTOCOL_RECEIPT`. The receipt is written by the
+protocol layer, which knows nothing about whether the work succeeded.
+
+**The simple broker declares what it cannot do.** `SimpleBrokerProfile` cannot be constructed
+claiming cluster support or durable acks, and refuses activation outside local/test — in a
+multi-node deployment it does not error, it delivers to whichever fraction of users is on the
+publishing node.
+
+**Unresolved user destinations are broadcast once and then dead-lettered.**
+`MultiNodeUserDestination` distinguishes a message that arrived *via* the broadcast from one that did
+not. Without that, every node rebroadcasts every unresolvable message on receipt.
+
+## Consequences
+
+- Enabling Advanced STOMP requires disabling the legacy channel. There is no migration path that
+  runs both; the exclusivity check makes that explicit at startup rather than at 3am.
+- A deployment must write its own catalog. There is deliberately no default: an empty one refuses
+  every frame and reads as a broken adapter, and a non-empty one publishes destinations nobody chose.
+- The relay's cost is one broker connection per authenticated session plus one system connection.
+  `brokerConnectionsFor` exists so this is computed before the first outage.
+- User-destination metrics are tagged with `UserDestinationAction`, never the destination — a user
+  destination contains a user identifier by construction.
+
+## Alternatives considered
+
+- **Extend the existing `stomp` package.** Rejected: it is Stable, and WS-ARCH-6 forbids a Stable
+  module naming an Advanced one. Making the legacy channel profile-driven would have required that
+  edge.
+- **One `advanced-stomp` module including the relay.** Rejected: the relay is a separate operational
+  decision and deserves to be refusable on its own.
+- **Allow undeclared destinations with a wildcard permission.** Rejected: the wildcard becomes the
+  default and the catalog becomes documentation.
@@ -0,0 +1,416 @@
+# GraphQL leaf public API surface — every public top-level type in src/main/java.
+# A public type in a single-jar leaf is reachable from every adopter's code, so
+# additions are reviewed rather than discovered. `api` and `spi` are the intended
+# external surface; the rest are candidates to become internal when this leaf is
+# split into capability artifacts.
+# Update only after review with:
+#   ./gradlew :adapter:inbound:graphql:updateGraphQlApiSurface -PapproveGraphQlApiSurfaceChange
+# types: 408
+dev.caskeleton.adapter.inbound.graphql.HealthGraphqlController
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlAdminPrincipal
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationAdminAuthorization
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationAdminDeniedException
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationAdminPort
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationAdminService
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationAudit
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationBlockCommand
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationRemovalGate
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationRemovalRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.admin.GraphQlPersistedOperationUsage
+dev.caskeleton.adapter.inbound.graphql.advanced.bootstrap.GraphQlAdvancedCapability
+dev.caskeleton.adapter.inbound.graphql.advanced.bootstrap.GraphQlAdvancedCapabilityDisabledException
+dev.caskeleton.adapter.inbound.graphql.advanced.bootstrap.GraphQlAdvancedCapabilityGrade
+dev.caskeleton.adapter.inbound.graphql.advanced.bootstrap.GraphQlAdvancedDependencyRules
+dev.caskeleton.adapter.inbound.graphql.advanced.bootstrap.GraphQlAdvancedFeatureFlags
+dev.caskeleton.adapter.inbound.graphql.advanced.bootstrap.GraphQlAdvancedModuleGuard
+dev.caskeleton.adapter.inbound.graphql.advanced.chaining.GraphQlChainedDataLoaderPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.chaining.GraphQlChainedDispatchConfigurer
+dev.caskeleton.adapter.inbound.graphql.advanced.chaining.GraphQlChainedLoaderMetrics
+dev.caskeleton.adapter.inbound.graphql.advanced.chaining.GraphQlDataLoaderCycleDetector
+dev.caskeleton.adapter.inbound.graphql.advanced.chaining.GraphQlDataLoaderDependencyCycleException
+dev.caskeleton.adapter.inbound.graphql.advanced.chaining.GraphQlDataLoaderDependencyGraph
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlClientOperationGenerator
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlCodegenBoundaryException
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlCodegenProfile
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlGeneratedCompatibilityGate
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlGeneratedSourceBoundary
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlOperationValidator
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlScalarMapping
+dev.caskeleton.adapter.inbound.graphql.advanced.codegen.GraphQlTransportTypeGenerator
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlFederationCompositionGate
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlFederationCompositionResult
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlFederationDeploymentOrder
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlFederationLatencyBudget
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlFederationReleaseEvidence
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlFederationReleaseRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlFederationUsageReport
+dev.caskeleton.adapter.inbound.graphql.advanced.composition.GraphQlSubgraphContract
+dev.caskeleton.adapter.inbound.graphql.advanced.federation.GraphQlFederationBatchResolver
+dev.caskeleton.adapter.inbound.graphql.advanced.federation.GraphQlFederationCapability
+dev.caskeleton.adapter.inbound.graphql.advanced.federation.GraphQlFederationEntityKey
+dev.caskeleton.adapter.inbound.graphql.advanced.federation.GraphQlFederationEntityResolver
+dev.caskeleton.adapter.inbound.graphql.advanced.federation.GraphQlFederationProperties
+dev.caskeleton.adapter.inbound.graphql.advanced.federation.GraphQlFederationRepresentationException
+dev.caskeleton.adapter.inbound.graphql.advanced.federation.GraphQlFederationSchemaFactory
+dev.caskeleton.adapter.inbound.graphql.advanced.get.GraphQlHttpDraftCompatibilityReport
+dev.caskeleton.adapter.inbound.graphql.advanced.get.GraphQlHttpGetCachePolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.get.GraphQlHttpGetCsrfPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.get.GraphQlHttpGetOperationPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.get.GraphQlHttpGetProfile
+dev.caskeleton.adapter.inbound.graphql.advanced.get.GraphQlHttpGetRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.get.GraphQlHttpGetRequestParser
+dev.caskeleton.adapter.inbound.graphql.advanced.incremental.GraphQlIncrementalCancellation
+dev.caskeleton.adapter.inbound.graphql.advanced.incremental.GraphQlIncrementalCompatibilityGate
+dev.caskeleton.adapter.inbound.graphql.advanced.incremental.GraphQlIncrementalDeliveryCapability
+dev.caskeleton.adapter.inbound.graphql.advanced.incremental.GraphQlIncrementalDeliveryProfile
+dev.caskeleton.adapter.inbound.graphql.advanced.incremental.GraphQlIncrementalDeliveryRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.incremental.GraphQlIncrementalPatch
+dev.caskeleton.adapter.inbound.graphql.advanced.incremental.GraphQlIncrementalTransportPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperation
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationConflictException
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationId
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationInterceptor
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationLookup
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationNotFoundException
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationRecordMapping
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationRegistry
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationRequest
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationStatus
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedOperationTransition
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.GraphQlPersistedPreparsedBridge
+dev.caskeleton.adapter.inbound.graphql.advanced.persisted.OperationalStoreGraphQlPersistedOperationRegistry
+dev.caskeleton.adapter.inbound.graphql.advanced.release.GraphQlAdvancedCompatibilityMatrix
+dev.caskeleton.adapter.inbound.graphql.advanced.release.GraphQlAdvancedPromotionDecision
+dev.caskeleton.adapter.inbound.graphql.advanced.release.GraphQlAdvancedReleaseEvidence
+dev.caskeleton.adapter.inbound.graphql.advanced.release.GraphQlAdvancedReleaseException
+dev.caskeleton.adapter.inbound.graphql.advanced.release.GraphQlAdvancedReleaseGate
+dev.caskeleton.adapter.inbound.graphql.advanced.release.GraphQlAdvancedRunbookIndex
+dev.caskeleton.adapter.inbound.graphql.advanced.release.GraphQlAdvancedSoakScenario
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlReplayAuthorization
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlReplayAuthorizationException
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlReplayGapException
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlReplayHistoryLostException
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlReplayPosition
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlReplaySource
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlSnapshotLiveHandoff
+dev.caskeleton.adapter.inbound.graphql.advanced.replay.GraphQlSubscriptionCursor
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketAdmission
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketAuthentication
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketCapability
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketErrorMapper
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketHandlerFactory
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketProperties
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketRoutePolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.rsocket.GraphQlRSocketRouteRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.security.GraphQlSubscriptionAuthorizationPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.security.GraphQlWebSocketAuthenticationException
+dev.caskeleton.adapter.inbound.graphql.advanced.security.GraphQlWebSocketAuthenticationInterceptor
+dev.caskeleton.adapter.inbound.graphql.advanced.security.GraphQlWebSocketCloseReason
+dev.caskeleton.adapter.inbound.graphql.advanced.security.GraphQlWebSocketCredentialExpiry
+dev.caskeleton.adapter.inbound.graphql.advanced.security.GraphQlWebSocketPrincipal
+dev.caskeleton.adapter.inbound.graphql.advanced.security.GraphQlWebSocketRevocationSignal
+dev.caskeleton.adapter.inbound.graphql.advanced.springdata.GraphQlRepositoryAllowlist
+dev.caskeleton.adapter.inbound.graphql.advanced.springdata.GraphQlRepositoryArgumentPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.springdata.GraphQlRepositoryExposure
+dev.caskeleton.adapter.inbound.graphql.advanced.springdata.GraphQlRepositoryExposureRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.springdata.GraphQlRepositoryExposureValidator
+dev.caskeleton.adapter.inbound.graphql.advanced.springdata.GraphQlRepositoryPaginationPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.springdata.GraphQlRepositoryProjectionPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.sse.GraphQlSseAdmission
+dev.caskeleton.adapter.inbound.graphql.advanced.sse.GraphQlSseConnectionPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.sse.GraphQlSseHandlerFactory
+dev.caskeleton.adapter.inbound.graphql.advanced.sse.GraphQlSseHeartbeat
+dev.caskeleton.adapter.inbound.graphql.advanced.sse.GraphQlSseProperties
+dev.caskeleton.adapter.inbound.graphql.advanced.sse.GraphQlSseRejectedException
+dev.caskeleton.adapter.inbound.graphql.advanced.sse.GraphQlSseTermination
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSlowConsumerPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionBufferPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionCancellation
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionContext
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionDispatcher
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionDrainCoordinator
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionDrainPhase
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionDrainingException
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionEvent
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionExecutionPolicy
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionLease
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionMetrics
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionOrderingProfile
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionSource
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionState
+dev.caskeleton.adapter.inbound.graphql.advanced.subscription.GraphQlSubscriptionTermination
+dev.caskeleton.adapter.inbound.graphql.advanced.websocket.GraphQlWebSocketAdmission
+dev.caskeleton.adapter.inbound.graphql.advanced.websocket.GraphQlWebSocketConnectionId
+dev.caskeleton.adapter.inbound.graphql.advanced.websocket.GraphQlWebSocketHandlerFactory
+dev.caskeleton.adapter.inbound.graphql.advanced.websocket.GraphQlWebSocketLifecycle
+dev.caskeleton.adapter.inbound.graphql.advanced.websocket.GraphQlWebSocketProperties
+dev.caskeleton.adapter.inbound.graphql.advanced.websocket.GraphQlWebSocketProtocol
+dev.caskeleton.adapter.inbound.graphql.advanced.websocket.GraphQlWebSocketProtocolException
+dev.caskeleton.adapter.inbound.graphql.api.GraphQlClientProfile
+dev.caskeleton.adapter.inbound.graphql.api.GraphQlClientProfileName
+dev.caskeleton.adapter.inbound.graphql.api.GraphQlOperationId
+dev.caskeleton.adapter.inbound.graphql.api.GraphQlOperationName
+dev.caskeleton.adapter.inbound.graphql.api.GraphQlSchemaCoordinate
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlAsyncReturnShape
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlControllerContractException
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlControllerInspector
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlControllerTransactionRule
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlInputTypePolicy
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlResolverBoundaryRules
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlReturnTypePolicy
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlTransportTypeRules
+dev.caskeleton.adapter.inbound.graphql.architecture.GraphQlTypeGraph
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlActivationEnvironmentPostProcessor
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlDeploymentMode
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlOffAutoConfigurationImportFilter
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlPlatformActuatorEndpoint
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlPlatformAutoConfiguration
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlPlatformConfigurationException
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlPlatformConfigurationReport
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlPlatformRuntime
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlPlatformSettings
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlPlatformStartupValidator
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlRetiredSafetyAxis
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlRootAutoConfiguration
+dev.caskeleton.adapter.inbound.graphql.autoconfigure.GraphQlRuntimeTransport
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlChangeKind
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlClientOwnerApproval
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlCompatibilityImpact
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlCompatibilityPolicy
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlCompatibilityReport
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlDeprecationGate
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlRemovalDecision
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlRemovalRequest
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlSchemaChange
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlSchemaComparator
+dev.caskeleton.adapter.inbound.graphql.compat.GraphQlSchemaUsage
+dev.caskeleton.adapter.inbound.graphql.context.ActorRef
+dev.caskeleton.adapter.inbound.graphql.context.GraphQlCommandAttribution
+dev.caskeleton.adapter.inbound.graphql.context.GraphQlDeadline
+dev.caskeleton.adapter.inbound.graphql.context.GraphQlIdentityFingerprinter
+dev.caskeleton.adapter.inbound.graphql.context.GraphQlRequestContext
+dev.caskeleton.adapter.inbound.graphql.context.TenantContext
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlComplexityCalculator
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlComplexityRejectedException
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlComplexityResult
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlCostCatalog
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlDocumentComplexityScorer
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlDocumentShape
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlDocumentShapeAnalyzer
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlFieldCostDescriptor
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlParserLimitPolicy
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlParserLimits
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlParserOptionsFactory
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlParserRejectedException
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlResolverWeight
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlResponseByteLimiter
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlResponseNodeCounter
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlRuntimeBudget
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlRuntimeBudgetExceededException
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlRuntimeBudgetTracker
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlStructuralLimitException
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlStructuralLimitPolicy
+dev.caskeleton.adapter.inbound.graphql.cost.GraphQlStructuralLimits
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchChunker
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchContext
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchErrorPolicy
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchExecutor
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchLoadException
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchObservation
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchPolicy
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchPolicyRegistry
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchResult
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchResultMapper
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchTimeoutException
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlBatchValue
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlDataLoaderFactory
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlDataLoaderName
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlDataLoaderRequestRegistry
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlMissingKeyException
+dev.caskeleton.adapter.inbound.graphql.dataloader.GraphQlMissingKeyPolicy
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlErrorCategory
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlErrorCode
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlErrorContext
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlExceptionResolver
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlFailureBoundary
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlInternalErrorMasker
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlNullabilityContract
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlRequestErrorMapper
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlSubscriptionExceptionResolver
+dev.caskeleton.adapter.inbound.graphql.error.GraphQlWireError
+dev.caskeleton.adapter.inbound.graphql.execution.BoundedPreparsedDocumentProvider
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlAnonymousOperationException
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlCancellation
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlDeadlinePropagator
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlExecutionPipeline
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlExecutionPipelineException
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlExecutionPipelineValidator
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlExecutionProfile
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlExecutionProfileException
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlExecutionProfileValidator
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlExecutionStage
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlOperationNameInterceptor
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlOperationNamePolicy
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlOperationSelection
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlPreparsedCacheKey
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlPreparsedCacheMetrics
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlPreparsedCachePolicy
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlRequestCancelledException
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlResolverBudget
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlResolverCatalog
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlResolverDescriptor
+dev.caskeleton.adapter.inbound.graphql.execution.GraphQlTimeoutPolicy
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlFetchProfile
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlFetchProfileClassifier
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlFetchProfileName
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlFetchProfileRegistry
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlFetchProfileRule
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlFetchProfileValidationException
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlSelectionCoordinate
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlSelectionSetView
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlSelectionSignature
+dev.caskeleton.adapter.inbound.graphql.fetch.GraphQlUnmappedSelectionException
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlAcceptHeader
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlExecutionOutcome
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlExtensionsPolicy
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpContractException
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpExecutor
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpOutcome
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpProfile
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpRequestEnvelope
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpResponse
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpResponseFactory
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpResponsePolicy
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlHttpStatusMapper
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlJsonStructurePolicy
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlJsonValues
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlMediaTypes
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlRequestEnvelopeValidator
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlRequestFormatException
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlRequestSize
+dev.caskeleton.adapter.inbound.graphql.http.GraphQlRequestTooLargeException
+dev.caskeleton.adapter.inbound.graphql.moduleboundary.GraphQlAdvancedModule
+dev.caskeleton.adapter.inbound.graphql.moduleboundary.GraphQlModuleBoundary
+dev.caskeleton.adapter.inbound.graphql.moduleboundary.GraphQlModulePurity
+dev.caskeleton.adapter.inbound.graphql.moduleboundary.GraphQlStableModule
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlBatchMutationItemResult
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlBusinessResult
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlCanonicalInput
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlExpectedVersion
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlIdempotencyConflictException
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlIdempotencyKey
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationContractException
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationContractValidator
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationCoordinate
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationFingerprint
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationIdempotencyContext
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationIdempotencyInterceptor
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationPayload
+dev.caskeleton.adapter.inbound.graphql.mutation.GraphQlMutationResultMapper
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlDataLoaderObservationConvention
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlMetricCardinalityPolicy
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlObservationContractException
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlObservationNames
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlOperationNameCardinality
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlProfilerAccessPolicy
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlRequestObservationConvention
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlResolverObservationConvention
+dev.caskeleton.adapter.inbound.graphql.observation.GraphQlSensitiveAttributeFilter
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlConnection
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlConnectionAssembler
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlConnectionException
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlConnectionPolicy
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlConnectionRequest
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorCodec
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorException
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorFraming
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorKeyRing
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorKeyset
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorPayload
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorScope
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlCursorVersion
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlEdge
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlKeysetWindow
+dev.caskeleton.adapter.inbound.graphql.pagination.GraphQlPageInfo
+dev.caskeleton.adapter.inbound.graphql.pagination.HmacGraphQlCursorCodec
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlClientPolicy
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlClientPolicyManifest
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlOperationCatalog
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlOperationPolicy
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlOperationType
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlPolicyViolation
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlUnknownClientProfileException
+dev.caskeleton.adapter.inbound.graphql.policy.GraphQlUnknownOperationException
+dev.caskeleton.adapter.inbound.graphql.policy.ResolverExecutionType
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlCompatibilityMatrix
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlFaultScenario
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlPerformanceScenario
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlReleaseEvidence
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlReleaseException
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlReleaseGate
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlReleaseOverride
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlReleaseReportWriter
+dev.caskeleton.adapter.inbound.graphql.release.GraphQlStableCapabilityManifest
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlBatchLoaderRegistrar
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlBlockingBridge
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlBlockingBridgeFullException
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlCostBudgetHandler
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlDataFetcherExceptionResolver
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlDocumentAuthorizationHandler
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlExecutionChain
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlExecutionContext
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlExecutionHandler
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlExecutionRequest
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlOperationSelectionHandler
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlPlatformInstrumentation
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlPlatformRejectionMapper
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlPlatformWebInterceptor
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlPreparsedDocumentAdapter
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlPrincipalResolver
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlRequestObservationConventionAdapter
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlWireErrorMapper
+dev.caskeleton.adapter.inbound.graphql.runtime.GraphQlWireErrors
+dev.caskeleton.adapter.inbound.graphql.runtime.servlet.GraphQlRequestBodyLimitFilter
+dev.caskeleton.adapter.inbound.graphql.scalar.BigDecimalScalar
+dev.caskeleton.adapter.inbound.graphql.scalar.DateScalar
+dev.caskeleton.adapter.inbound.graphql.scalar.GraphQlDecimalBounds
+dev.caskeleton.adapter.inbound.graphql.scalar.GraphQlScalarWiringConfigurer
+dev.caskeleton.adapter.inbound.graphql.scalar.InstantScalar
+dev.caskeleton.adapter.inbound.graphql.scalar.LongScalar
+dev.caskeleton.adapter.inbound.graphql.scalar.UuidScalar
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlContractVersion
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlMappingInspectionGate
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlMappingIssue
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlMappingPolicy
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlOneOfInputValidator
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlOneOfPolicy
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlOneOfSchemaGate
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlOneOfViolationException
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlScalarDefinition
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlScalarManifest
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlScalarPolicy
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaAssembler
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaAssemblyException
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaAssemblyResult
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaContract
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaHash
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaMappingException
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaOwnership
+dev.caskeleton.adapter.inbound.graphql.schema.GraphQlSchemaResource
+dev.caskeleton.adapter.inbound.graphql.security.ApplicationObjectAuthorization
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlAuthenticatedPrincipal
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlAuthenticationContextFactory
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlAuthenticationException
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlAuthorizationDecision
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlAuthorizationDeniedException
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlAuthorizationInterceptor
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlAuthorizationPolicy
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlBatchContext
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlClientProfileResolver
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlContextCleanup
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlContextPropagator
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlObjectAuthorizationPort
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlTenantIsolationException
+dev.caskeleton.adapter.inbound.graphql.security.GraphQlTenantIsolationPolicy
@@ -0,0 +1,346 @@
+# JPA persistence leaf public API surface — every public top-level type in src/main/java.
+# A public type in a single-jar leaf is reachable from every adopter's code, so
+# additions are reviewed rather than discovered. `api` is the intended external
+# surface; the rest is implementation that has not been moved under an internal
+# root yet.
+# Update only after review with:
+#   ./gradlew :adapter:outbound:persistence-jpa:updateJpaApiSurface -PapproveJpaApiSurfaceChange
+# types: 338
+dev.caskeleton.adapter.outbound.persistence.api.PersistenceOperationName
+dev.caskeleton.adapter.outbound.persistence.api.capability.CapabilitySupport
+dev.caskeleton.adapter.outbound.persistence.api.capability.JpaCapability
+dev.caskeleton.adapter.outbound.persistence.api.capability.SupportLevel
+dev.caskeleton.adapter.outbound.persistence.api.error.CheckConstraintViolationException
+dev.caskeleton.adapter.outbound.persistence.api.error.ConnectionUnavailableException
+dev.caskeleton.adapter.outbound.persistence.api.error.ConstraintCode
+dev.caskeleton.adapter.outbound.persistence.api.error.ConstraintViolationDetails
+dev.caskeleton.adapter.outbound.persistence.api.error.DataCorruptionException
+dev.caskeleton.adapter.outbound.persistence.api.error.DeadlockDetectedException
+dev.caskeleton.adapter.outbound.persistence.api.error.FailureCategory
+dev.caskeleton.adapter.outbound.persistence.api.error.ForeignKeyViolationException
+dev.caskeleton.adapter.outbound.persistence.api.error.JpaEntityNotFoundException
+dev.caskeleton.adapter.outbound.persistence.api.error.JpaFailureContext
+dev.caskeleton.adapter.outbound.persistence.api.error.JpaPersistenceException
+dev.caskeleton.adapter.outbound.persistence.api.error.NotNullConstraintViolationException
+dev.caskeleton.adapter.outbound.persistence.api.error.OptimisticConflictException
+dev.caskeleton.adapter.outbound.persistence.api.error.PessimisticLockTimeoutException
+dev.caskeleton.adapter.outbound.persistence.api.error.QueryTimeoutException
+dev.caskeleton.adapter.outbound.persistence.api.error.SchemaMismatchException
+dev.caskeleton.adapter.outbound.persistence.api.error.SerializationFailureException
+dev.caskeleton.adapter.outbound.persistence.api.error.SqlExceptionSqlStateResolver
+dev.caskeleton.adapter.outbound.persistence.api.error.SqlStateResolver
+dev.caskeleton.adapter.outbound.persistence.api.error.TransactionCompletionUnknownException
+dev.caskeleton.adapter.outbound.persistence.api.error.TransactionTimeoutException
+dev.caskeleton.adapter.outbound.persistence.api.error.UniqueConstraintViolationException
+dev.caskeleton.adapter.outbound.persistence.api.error.VendorFailureTranslator
+dev.caskeleton.adapter.outbound.persistence.api.query.CursorCodec
+dev.caskeleton.adapter.outbound.persistence.api.query.CursorPayloadCodec
+dev.caskeleton.adapter.outbound.persistence.api.query.KeysetPageRequest
+dev.caskeleton.adapter.outbound.persistence.api.query.KeysetSlice
+dev.caskeleton.adapter.outbound.persistence.api.query.NoopQueryObservation
+dev.caskeleton.adapter.outbound.persistence.api.query.QueryName
+dev.caskeleton.adapter.outbound.persistence.api.query.QueryObservation
+dev.caskeleton.adapter.outbound.persistence.api.query.QueryScope
+dev.caskeleton.adapter.outbound.persistence.api.query.SignedJsonCursorCodec
+dev.caskeleton.adapter.outbound.persistence.api.query.SortDirection
+dev.caskeleton.adapter.outbound.persistence.api.transaction.IsolationLevel
+dev.caskeleton.adapter.outbound.persistence.api.transaction.JitterMode
+dev.caskeleton.adapter.outbound.persistence.api.transaction.JpaRetryPolicy
+dev.caskeleton.adapter.outbound.persistence.api.transaction.JpaTransactionExecutor
+dev.caskeleton.adapter.outbound.persistence.api.transaction.PropagationMode
+dev.caskeleton.adapter.outbound.persistence.api.transaction.RetryDecision
+dev.caskeleton.adapter.outbound.persistence.api.transaction.RetryDisposition
+dev.caskeleton.adapter.outbound.persistence.api.transaction.RetryEventListener
+dev.caskeleton.adapter.outbound.persistence.api.transaction.RetryProfile
+dev.caskeleton.adapter.outbound.persistence.api.transaction.TransactionAttempt
+dev.caskeleton.adapter.outbound.persistence.api.transaction.TransactionCompletionEvidence
+dev.caskeleton.adapter.outbound.persistence.api.transaction.TransactionProfile
+dev.caskeleton.adapter.outbound.persistence.audit.AuditContextPort
+dev.caskeleton.adapter.outbound.persistence.audit.AuditableEntity
+dev.caskeleton.adapter.outbound.persistence.audit.DomainContextAuditContextPort
+dev.caskeleton.adapter.outbound.persistence.auditing.AuditMetadata
+dev.caskeleton.adapter.outbound.persistence.auditing.JpaAuditingConfiguration
+dev.caskeleton.adapter.outbound.persistence.auditing.JpaAuditorProvider
+dev.caskeleton.adapter.outbound.persistence.cache.CacheConcurrencyStrategy
+dev.caskeleton.adapter.outbound.persistence.cache.CacheRegionCatalog
+dev.caskeleton.adapter.outbound.persistence.cache.HibernateCacheGuard
+dev.caskeleton.adapter.outbound.persistence.cache.HibernateCachePolicy
+dev.caskeleton.adapter.outbound.persistence.cache.HibernateCacheSettings
+dev.caskeleton.adapter.outbound.persistence.config.JpaAdapterComponentsConfig
+dev.caskeleton.adapter.outbound.persistence.config.PersistenceJpaConfig
+dev.caskeleton.adapter.outbound.persistence.config.PersistenceVendorSettings
+dev.caskeleton.adapter.outbound.persistence.envers.EntityRevision
+dev.caskeleton.adapter.outbound.persistence.envers.EnversConfigurationGuard
+dev.caskeleton.adapter.outbound.persistence.envers.EnversHistoryPolicy
+dev.caskeleton.adapter.outbound.persistence.envers.EnversHistoryReader
+dev.caskeleton.adapter.outbound.persistence.envers.EnversRevisionMetadata
+dev.caskeleton.adapter.outbound.persistence.envers.HibernateEnversHistoryReader
+dev.caskeleton.adapter.outbound.persistence.experimental.ExperimentalFeature
+dev.caskeleton.adapter.outbound.persistence.experimental.ExperimentalFeatureGate
+dev.caskeleton.adapter.outbound.persistence.experimental.database.TenantDataSourceLifecycle
+dev.caskeleton.adapter.outbound.persistence.experimental.database.TenantDataSourceRegistry
+dev.caskeleton.adapter.outbound.persistence.experimental.database.TenantEntityManagerFactoryRegistry
+dev.caskeleton.adapter.outbound.persistence.experimental.database.TenantPoolBudget
+dev.caskeleton.adapter.outbound.persistence.experimental.next.CompatibilityLane
+dev.caskeleton.adapter.outbound.persistence.experimental.next.ExperimentalPromotionGate
+dev.caskeleton.adapter.outbound.persistence.experimental.next.HibernateCompatibilityPolicy
+dev.caskeleton.adapter.outbound.persistence.experimental.next.PromotionDecision
+dev.caskeleton.adapter.outbound.persistence.experimental.next.PromotionEvidence
+dev.caskeleton.adapter.outbound.persistence.experimental.replica.ConsistencyAwareDataSourceRouter
+dev.caskeleton.adapter.outbound.persistence.experimental.replica.ConsistencyToken
+dev.caskeleton.adapter.outbound.persistence.experimental.replica.ReadConsistency
+dev.caskeleton.adapter.outbound.persistence.experimental.replica.ReplicaLagMonitor
+dev.caskeleton.adapter.outbound.persistence.experimental.replica.ReplicaRoutingDecision
+dev.caskeleton.adapter.outbound.persistence.experimental.replica.ReplicaTarget
+dev.caskeleton.adapter.outbound.persistence.experimental.replica.TransactionContext
+dev.caskeleton.adapter.outbound.persistence.experimental.rls.RlsAdminBypassToken
+dev.caskeleton.adapter.outbound.persistence.experimental.rls.RlsPolicyVerifier
+dev.caskeleton.adapter.outbound.persistence.experimental.rls.RlsTenantSessionBinder
+dev.caskeleton.adapter.outbound.persistence.experimental.schema.SchemaMultiTenantConnectionProvider
+dev.caskeleton.adapter.outbound.persistence.experimental.schema.SchemaTenantMigrationOrchestrator
+dev.caskeleton.adapter.outbound.persistence.experimental.schema.SchemaTenantRegistry
+dev.caskeleton.adapter.outbound.persistence.experimental.schema.TenantMigrationStatus
+dev.caskeleton.adapter.outbound.persistence.experimental.tenant.TenantAwareRepositoryGuard
+dev.caskeleton.adapter.outbound.persistence.experimental.tenant.TenantContext
+dev.caskeleton.adapter.outbound.persistence.experimental.tenant.TenantEntityListenerGuard
+dev.caskeleton.adapter.outbound.persistence.experimental.tenant.TenantId
+dev.caskeleton.adapter.outbound.persistence.failure.PersistenceExceptionTranslator
+dev.caskeleton.adapter.outbound.persistence.failure.SqlStateErrorMapping
+dev.caskeleton.adapter.outbound.persistence.failure.StandardSqlStateErrorMapping
+dev.caskeleton.adapter.outbound.persistence.fileserver.FileserverJpaPersistenceConfig
+dev.caskeleton.adapter.outbound.persistence.fileserver.FileserverSchemaActivation
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaCleanupQueue
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaContentReferenceLedger
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaFileMetadataStore
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaFileQuotaService
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaQuotaCommitGateway
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaQuotaReclaimGateway
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaRecoveryQueue
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaStagingUploadLocator
+dev.caskeleton.adapter.outbound.persistence.fileserver.JpaUploadSessionStore
+dev.caskeleton.adapter.outbound.persistence.fileserver.entity.CleanupItemEntity
+dev.caskeleton.adapter.outbound.persistence.fileserver.entity.FileEntity
+dev.caskeleton.adapter.outbound.persistence.fileserver.entity.QuotaReservationEntity
+dev.caskeleton.adapter.outbound.persistence.fileserver.entity.RecoveryItemEntity
+dev.caskeleton.adapter.outbound.persistence.fileserver.entity.UploadSessionEntity
+dev.caskeleton.adapter.outbound.persistence.fileserver.entity.VerificationResultEntity
+dev.caskeleton.adapter.outbound.persistence.fileserver.repository.FileTransitionRepository
+dev.caskeleton.adapter.outbound.persistence.fileserver.repository.FileserverCleanupRepository
+dev.caskeleton.adapter.outbound.persistence.fileserver.repository.FileserverQuotaRepository
+dev.caskeleton.adapter.outbound.persistence.fileserver.repository.FileserverRecoveryRepository
+dev.caskeleton.adapter.outbound.persistence.fileserver.repository.JpaFileRepository
+dev.caskeleton.adapter.outbound.persistence.fileserver.repository.JpaUploadSessionRepository
+dev.caskeleton.adapter.outbound.persistence.fileserver.repository.UploadLeaseRepository
+dev.caskeleton.adapter.outbound.persistence.h2.H2IdempotencyClaimRepository
+dev.caskeleton.adapter.outbound.persistence.h2.H2LocalTimeoutConfigurer
+dev.caskeleton.adapter.outbound.persistence.h2.H2OutboxClaimRepository
+dev.caskeleton.adapter.outbound.persistence.h2.H2PersistenceConfig
+dev.caskeleton.adapter.outbound.persistence.h2.H2SqlStateErrorMapping
+dev.caskeleton.adapter.outbound.persistence.hibernate.HibernateProviderPolicy
+dev.caskeleton.adapter.outbound.persistence.hibernate.HibernateStatisticsCollector
+dev.caskeleton.adapter.outbound.persistence.hibernate.HibernateStatisticsSnapshot
+dev.caskeleton.adapter.outbound.persistence.hibernate.JdbcBatchCounter
+dev.caskeleton.adapter.outbound.persistence.hibernate.NamedStatementInspector
+dev.caskeleton.adapter.outbound.persistence.hibernate.QueryNameContext
+dev.caskeleton.adapter.outbound.persistence.hibernate.batch.BatchExecutionResult
+dev.caskeleton.adapter.outbound.persistence.hibernate.batch.HibernateBatchConfigurationGuard
+dev.caskeleton.adapter.outbound.persistence.hibernate.batch.HibernateJpaBatchExecutor
+dev.caskeleton.adapter.outbound.persistence.hibernate.batch.JpaBatchExecutor
+dev.caskeleton.adapter.outbound.persistence.hibernate.batch.JpaBatchProfile
+dev.caskeleton.adapter.outbound.persistence.hibernate.batch.JpaBatchProfileRegistry
+dev.caskeleton.adapter.outbound.persistence.hibernate.bulk.AffectedRowsExpectation
+dev.caskeleton.adapter.outbound.persistence.hibernate.bulk.BulkDmlExecutor
+dev.caskeleton.adapter.outbound.persistence.hibernate.bulk.BulkDmlResult
+dev.caskeleton.adapter.outbound.persistence.hibernate.bulk.BulkOperationName
+dev.caskeleton.adapter.outbound.persistence.hibernate.bulk.HibernateBulkDmlExecutor
+dev.caskeleton.adapter.outbound.persistence.hibernate.stateless.HibernateStatelessSessionRunner
+dev.caskeleton.adapter.outbound.persistence.hibernate.stateless.StatelessRowCapExceededException
+dev.caskeleton.adapter.outbound.persistence.hibernate.stateless.StatelessSessionRunner
+dev.caskeleton.adapter.outbound.persistence.hibernate.stateless.StatelessWorkName
+dev.caskeleton.adapter.outbound.persistence.hibernate.stateless.StatelessWorkResult
+dev.caskeleton.adapter.outbound.persistence.idempotency.IdempotencyClaimRepository
+dev.caskeleton.adapter.outbound.persistence.idempotency.IdempotencyReaper
+dev.caskeleton.adapter.outbound.persistence.idempotency.IdempotencyRecordJpaRepository
+dev.caskeleton.adapter.outbound.persistence.idempotency.IdempotencyResponseObjectStore
+dev.caskeleton.adapter.outbound.persistence.idempotency.IdempotencyStoreAdapter
+dev.caskeleton.adapter.outbound.persistence.idempotency.entity.IdempotencyRecordEntity
+dev.caskeleton.adapter.outbound.persistence.idempotency.mapper.IdempotencyRecordEntityMapper
+dev.caskeleton.adapter.outbound.persistence.liveevent.JpaLiveEventReplayAdapter
+dev.caskeleton.adapter.outbound.persistence.liveevent.LiveEventJpaRepository
+dev.caskeleton.adapter.outbound.persistence.liveevent.entity.LiveEventEntity
+dev.caskeleton.adapter.outbound.persistence.lock.DistributedLockPersistenceConfig
+dev.caskeleton.adapter.outbound.persistence.lock.LockRegistryDistributedLockAdapter
+dev.caskeleton.adapter.outbound.persistence.lock.LockSettings
+dev.caskeleton.adapter.outbound.persistence.migration.ConcurrentIndexMigrationInspector
+dev.caskeleton.adapter.outbound.persistence.migration.FailedConcurrentIndexRecovery
+dev.caskeleton.adapter.outbound.persistence.migration.FlywaySchemaPolicy
+dev.caskeleton.adapter.outbound.persistence.migration.FlywayValidationGate
+dev.caskeleton.adapter.outbound.persistence.migration.MigrationResource
+dev.caskeleton.adapter.outbound.persistence.migration.NonTransactionalMigrationPolicy
+dev.caskeleton.adapter.outbound.persistence.migration.SchemaManagementMode
+dev.caskeleton.adapter.outbound.persistence.migration.SchemaVersionSnapshot
+dev.caskeleton.adapter.outbound.persistence.notification.NotificationJpaPersistenceConfig
+dev.caskeleton.adapter.outbound.persistence.notification.NotificationSchemaActivation
+dev.caskeleton.adapter.outbound.persistence.notification.NotificationSchemaStream
+dev.caskeleton.adapter.outbound.persistence.notification.configuration.NotificationJpaPersistenceFacade
+dev.caskeleton.adapter.outbound.persistence.notification.crypto.DirectAeadNotificationPayloadCrypto
+dev.caskeleton.adapter.outbound.persistence.notification.crypto.NotificationCiphertext
+dev.caskeleton.adapter.outbound.persistence.notification.crypto.NotificationCryptoException
+dev.caskeleton.adapter.outbound.persistence.notification.crypto.NotificationHmacDigester
+dev.caskeleton.adapter.outbound.persistence.notification.crypto.NotificationKeyMaterialHandle
+dev.caskeleton.adapter.outbound.persistence.notification.crypto.NotificationKeyMaterialProvider
+dev.caskeleton.adapter.outbound.persistence.notification.platform.AdminAuditEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.AdminAuditJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.ConsentEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.ConsentJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.ContactPointEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.ContactPointJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.DeduplicationClaimEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.DeduplicationClaimJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.DeliveryAttemptEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.DeliveryAttemptJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JdbcNotificationServingState
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JdbcReconciliationJobStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaAdminOperationStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaContactPointStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaDeliveryAttemptStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaNotificationRequestStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaNotificationSideEffectStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaPolicyStores
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaProviderEventLedger
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaRecipientDeliveryStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaRecipientLeaseStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaSuppressionStore
+dev.caskeleton.adapter.outbound.persistence.notification.platform.JpaTemplateRegistry
+dev.caskeleton.adapter.outbound.persistence.notification.platform.NotificationRecordMapper
+dev.caskeleton.adapter.outbound.persistence.notification.platform.NotificationRequestEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.NotificationRequestJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.PreferenceEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.PreferenceJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.ProviderEventEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.ProviderEventJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.RecipientClaimSql
+dev.caskeleton.adapter.outbound.persistence.notification.platform.RecipientDeliveryEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.RecipientDeliveryJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.SuppressionEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.SuppressionJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.TemplateVersionEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.TemplateVersionJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.TenantBoundRepositoryGuard
+dev.caskeleton.adapter.outbound.persistence.notification.platform.inbox.InboxCommitEventPublisher
+dev.caskeleton.adapter.outbound.persistence.notification.platform.inbox.InboxItemEntity
+dev.caskeleton.adapter.outbound.persistence.notification.platform.inbox.InboxItemJpaRepository
+dev.caskeleton.adapter.outbound.persistence.notification.platform.inbox.InboxOutboxRecordFactory
+dev.caskeleton.adapter.outbound.persistence.notification.platform.inbox.JpaNotificationInbox
+dev.caskeleton.adapter.outbound.persistence.observation.JpaMetricTags
+dev.caskeleton.adapter.outbound.persistence.observation.JpaRetryObservation
+dev.caskeleton.adapter.outbound.persistence.observation.JpaTransactionObservation
+dev.caskeleton.adapter.outbound.persistence.observation.LowCardinality
+dev.caskeleton.adapter.outbound.persistence.observation.MicrometerQueryObservation
+dev.caskeleton.adapter.outbound.persistence.observation.SqlDiagnosticRedactor
+dev.caskeleton.adapter.outbound.persistence.operation.DurableOperationJpaRepository
+dev.caskeleton.adapter.outbound.persistence.operation.DurableOperationStoreAdapter
+dev.caskeleton.adapter.outbound.persistence.operation.entity.DurableOperationEntity
+dev.caskeleton.adapter.outbound.persistence.outbox.OutboxClaimRepository
+dev.caskeleton.adapter.outbound.persistence.outbox.OutboxEventJpaRepository
+dev.caskeleton.adapter.outbound.persistence.outbox.OutboxReaper
+dev.caskeleton.adapter.outbound.persistence.outbox.OutboxStoreAdapter
+dev.caskeleton.adapter.outbound.persistence.outbox.entity.OutboxEventEntity
+dev.caskeleton.adapter.outbound.persistence.postgresql.PostgreSqlIdempotencyClaimRepository
+dev.caskeleton.adapter.outbound.persistence.postgresql.PostgreSqlLocalTimeoutConfigurer
+dev.caskeleton.adapter.outbound.persistence.postgresql.PostgreSqlOutboxClaimRepository
+dev.caskeleton.adapter.outbound.persistence.postgresql.PostgreSqlPersistenceConfig
+dev.caskeleton.adapter.outbound.persistence.postgresql.PostgreSqlSqlStateErrorMapping
+dev.caskeleton.adapter.outbound.persistence.postgresql.array.PostgreSqlArraySupport
+dev.caskeleton.adapter.outbound.persistence.postgresql.constraint.PostgreSqlConstraintCatalog
+dev.caskeleton.adapter.outbound.persistence.postgresql.constraint.PostgreSqlConstraintViolationTranslator
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.BoundedCopyInputStream
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.CopyAdminCapability
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.CopyFormat
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.CopyLimits
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.CopyOperationName
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.CopyResult
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.PostgreSqlCopyLoader
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.RegisteredCopyStatement
+dev.caskeleton.adapter.outbound.persistence.postgresql.copy.RegisteredPostgreSqlCopyLoader
+dev.caskeleton.adapter.outbound.persistence.postgresql.error.ConstraintCatalog
+dev.caskeleton.adapter.outbound.persistence.postgresql.error.PostgreSqlExceptionTranslator
+dev.caskeleton.adapter.outbound.persistence.postgresql.error.PostgreSqlFailureClassifier
+dev.caskeleton.adapter.outbound.persistence.postgresql.error.PostgreSqlServerErrorFields
+dev.caskeleton.adapter.outbound.persistence.postgresql.error.PostgreSqlState
+dev.caskeleton.adapter.outbound.persistence.postgresql.idempotency.PostgreSqlOwnerSafeIdempotencyStore
+dev.caskeleton.adapter.outbound.persistence.postgresql.inbox.PostgreSqlSameStoreInboxAdapter
+dev.caskeleton.adapter.outbound.persistence.postgresql.json.JsonDocument
+dev.caskeleton.adapter.outbound.persistence.postgresql.json.JsonDocumentCodec
+dev.caskeleton.adapter.outbound.persistence.postgresql.json.JsonPathName
+dev.caskeleton.adapter.outbound.persistence.postgresql.json.PostgreSqlJsonQuerySupport
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.LockWaitObservation
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.PostgreSqlLockExceptionTranslator
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.PostgreSqlLockOptions
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.PostgreSqlWorkClaimExecutor
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.WorkClaim
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.WorkClaimExecutor
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.WorkQueueDefinition
+dev.caskeleton.adapter.outbound.persistence.postgresql.lock.WorkQueueName
+dev.caskeleton.adapter.outbound.persistence.postgresql.outbox.PostgreSqlImmutableOutboxAppendAdapter
+dev.caskeleton.adapter.outbound.persistence.postgresql.outbox.PostgreSqlPollingDeliveryAdapter
+dev.caskeleton.adapter.outbound.persistence.postgresql.range.PgRange
+dev.caskeleton.adapter.outbound.persistence.postgresql.range.PgRangeCodec
+dev.caskeleton.adapter.outbound.persistence.postgresql.range.PgRangeJdbcType
+dev.caskeleton.adapter.outbound.persistence.postgresql.range.PostgreSqlRangeQuerySupport
+dev.caskeleton.adapter.outbound.persistence.postgresql.write.NativeWriteName
+dev.caskeleton.adapter.outbound.persistence.postgresql.write.PostgreSqlUpsertExecutor
+dev.caskeleton.adapter.outbound.persistence.postgresql.write.RegisteredPostgreSqlUpsertExecutor
+dev.caskeleton.adapter.outbound.persistence.postgresql.write.RegisteredUpsertStatement
+dev.caskeleton.adapter.outbound.persistence.postgresql.write.UpsertConflictTarget
+dev.caskeleton.adapter.outbound.persistence.postgresql.write.UpsertResult
+dev.caskeleton.adapter.outbound.persistence.postgresql.write.WriteDisposition
+dev.caskeleton.adapter.outbound.persistence.querydsl.PredicatePolicy
+dev.caskeleton.adapter.outbound.persistence.querydsl.QueryPage
+dev.caskeleton.adapter.outbound.persistence.querydsl.QuerydslJpaSupport
+dev.caskeleton.adapter.outbound.persistence.security.DatabasePrivilegeReport
+dev.caskeleton.adapter.outbound.persistence.security.DatabaseRolePolicy
+dev.caskeleton.adapter.outbound.persistence.security.PostgreSqlRuntimeRoleVerifier
+dev.caskeleton.adapter.outbound.persistence.security.SearchPathPolicy
+dev.caskeleton.adapter.outbound.persistence.springdata.EntityGraphCatalog
+dev.caskeleton.adapter.outbound.persistence.springdata.EntityManagerAccess
+dev.caskeleton.adapter.outbound.persistence.springdata.FetchPlanApplier
+dev.caskeleton.adapter.outbound.persistence.springdata.FetchPlanName
+dev.caskeleton.adapter.outbound.persistence.springdata.JpaKeysetQuerySupport
+dev.caskeleton.adapter.outbound.persistence.springdata.JpaRepositoryFragmentSupport
+dev.caskeleton.adapter.outbound.persistence.springdata.JpaStreamExecutor
+dev.caskeleton.adapter.outbound.persistence.springdata.JpaStreamScope
+dev.caskeleton.adapter.outbound.persistence.springdata.KeysetPredicateBuilder
+dev.caskeleton.adapter.outbound.persistence.springdata.KeysetSliceAssembler
+dev.caskeleton.adapter.outbound.persistence.springdata.KeysetTerm
+dev.caskeleton.adapter.outbound.persistence.springdata.RegisteredQuery
+dev.caskeleton.adapter.outbound.persistence.springdata.SafeSortField
+dev.caskeleton.adapter.outbound.persistence.springdata.SafeSortMapper
+dev.caskeleton.adapter.outbound.persistence.springdata.SafeSortRegistry
+dev.caskeleton.adapter.outbound.persistence.springdata.ScrollPolicy
+dev.caskeleton.adapter.outbound.persistence.springdata.SpecificationPolicy
+dev.caskeleton.adapter.outbound.persistence.transaction.BackoffCalculator
+dev.caskeleton.adapter.outbound.persistence.transaction.CommitFailureClassifier
+dev.caskeleton.adapter.outbound.persistence.transaction.CompletionUnknownRecord
+dev.caskeleton.adapter.outbound.persistence.transaction.CompletionUnknownRecorder
+dev.caskeleton.adapter.outbound.persistence.transaction.DefaultJpaRetryPolicy
+dev.caskeleton.adapter.outbound.persistence.transaction.EffectiveTransactionTimeouts
+dev.caskeleton.adapter.outbound.persistence.transaction.EvidenceAwareJpaTransactionManager
+dev.caskeleton.adapter.outbound.persistence.transaction.FullTransactionRetryCoordinator
+dev.caskeleton.adapter.outbound.persistence.transaction.JpaTransactionConfig
+dev.caskeleton.adapter.outbound.persistence.transaction.JpaTransactionSettings
+dev.caskeleton.adapter.outbound.persistence.transaction.OptimisticConflictTranslator
+dev.caskeleton.adapter.outbound.persistence.transaction.PersistenceFailureTranslatorChain
+dev.caskeleton.adapter.outbound.persistence.transaction.RetryBudget
+dev.caskeleton.adapter.outbound.persistence.transaction.RetrySleeper
+dev.caskeleton.adapter.outbound.persistence.transaction.SpringJpaTransactionExecutor
+dev.caskeleton.adapter.outbound.persistence.transaction.SpringTransactionPort
+dev.caskeleton.adapter.outbound.persistence.transaction.ThreadRetrySleeper
+dev.caskeleton.adapter.outbound.persistence.transaction.TransactionDefinitionMapper
+dev.caskeleton.adapter.outbound.persistence.transaction.TransactionEvidenceContext
+dev.caskeleton.adapter.outbound.persistence.transaction.TransactionEvidenceFrame
+dev.caskeleton.adapter.outbound.persistence.transaction.TransactionEvidenceScope
+dev.caskeleton.adapter.outbound.persistence.transaction.TransactionLocalTimeoutConfigurer
+dev.caskeleton.adapter.outbound.persistence.transaction.TransactionProfileRegistry
+dev.caskeleton.adapter.outbound.persistence.transaction.TransactionStartBudget
+dev.caskeleton.adapter.outbound.persistence.transaction.UnknownOperation
@@ -0,0 +1,354 @@
+# MongoDB leaf public API surface — every public top-level type in src/main/java.
+# A public type in a single-jar leaf is reachable from every adopter's code, so
+# additions are reviewed rather than discovered. `api` is the intended external
+# surface; the rest is implementation that has not been moved under an internal
+# root yet.
+# Update only after review with:
+#   ./gradlew :adapter:outbound:persistence-mongo:updateMongoApiSurface -PapproveMongoApiSurfaceChange
+# types: 346
+dev.caskeleton.adapter.outbound.mongo.MongoOptInAutoConfigurationImportFilter
+dev.caskeleton.adapter.outbound.mongo.MongoPersistenceConfig
+dev.caskeleton.adapter.outbound.mongo.MongoPersistenceSettings
+dev.caskeleton.adapter.outbound.mongo.MongoRootAutoConfiguration
+dev.caskeleton.adapter.outbound.mongo.advanced.MongoAdvancedCapabilityFlags
+dev.caskeleton.adapter.outbound.mongo.advanced.MongoAdvancedCapabilityGuard
+dev.caskeleton.adapter.outbound.mongo.advanced.MongoAdvancedEntryPoint
+dev.caskeleton.adapter.outbound.mongo.advanced.MongoAdvancedPolicy
+dev.caskeleton.adapter.outbound.mongo.advanced.MongoAdvancedPromotionEvidence
+dev.caskeleton.adapter.outbound.mongo.advanced.MongoAdvancedPromotionGate
+dev.caskeleton.adapter.outbound.mongo.advanced.autoconfigure.MongoAdvancedConfiguration
+dev.caskeleton.adapter.outbound.mongo.advanced.autoconfigure.MongoAdvancedSettings
+dev.caskeleton.adapter.outbound.mongo.advanced.bridge.MongoBridgeCheckpointPolicy
+dev.caskeleton.adapter.outbound.mongo.advanced.bridge.MongoBridgeOutboxPolicy
+dev.caskeleton.adapter.outbound.mongo.advanced.bridge.MongoChangeMessagingBridge
+dev.caskeleton.adapter.outbound.mongo.advanced.bridge.MongoChangeToIntegrationEventMapper
+dev.caskeleton.adapter.outbound.mongo.advanced.bridge.MongoIntegrationEventEnvelope
+dev.caskeleton.adapter.outbound.mongo.advanced.bridge.MongoIntegrationEventPublisher
+dev.caskeleton.adapter.outbound.mongo.advanced.bridge.MongoPublishResult
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.csfle.MongoCsfleClientFactory
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.csfle.MongoCsfleFieldPolicy
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.csfle.MongoCsfleMode
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.csfle.MongoCsfleProfile
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.csfle.MongoDataKeyResolver
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.qe.MongoEncryptedFieldDescriptor
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.qe.MongoEncryptionMetadataOwnership
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.qe.MongoQueryShape
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.qe.MongoQueryShapeSupport
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.qe.MongoQueryableEncryptionCollectionManager
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.qe.MongoQueryableEncryptionProfile
+dev.caskeleton.adapter.outbound.mongo.advanced.encryption.qe.MongoQueryableEncryptionQueryType
+dev.caskeleton.adapter.outbound.mongo.advanced.gridfs.MongoGridFsCompatibilityReader
+dev.caskeleton.adapter.outbound.mongo.advanced.gridfs.MongoGridFsMigrationCheckpoint
+dev.caskeleton.adapter.outbound.mongo.advanced.gridfs.MongoGridFsMigrationJob
+dev.caskeleton.adapter.outbound.mongo.advanced.gridfs.MongoGridFsObjectReference
+dev.caskeleton.adapter.outbound.mongo.advanced.search.MongoSearchIndexDescriptor
+dev.caskeleton.adapter.outbound.mongo.advanced.search.MongoSearchIndexState
+dev.caskeleton.adapter.outbound.mongo.advanced.search.MongoSearchOperations
+dev.caskeleton.adapter.outbound.mongo.advanced.search.MongoSearchQuery
+dev.caskeleton.adapter.outbound.mongo.advanced.search.MongoSearchReadinessGate
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.MongoRoutingClassification
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.ShardAwareQueryValidator
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.ShardKeyDescriptor
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.ShardKeyPart
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.ShardStrategy
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.admin.MongoShardingAdminGateway
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.admin.ReshardApproval
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.admin.ShardKeyAnalyzer
+dev.caskeleton.adapter.outbound.mongo.advanced.sharding.admin.ShardKeyReadinessReport
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.database.MongoTenantClientRegistry
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.database.MongoTenantDatabaseResolver
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.database.MongoTenantLifecyclePolicy
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.database.MongoTenantMigrationCheckpointStore
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.database.MongoTenantMigrationCoordinator
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.shared.MongoTenantContext
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.shared.MongoTenantManifestValidator
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.shared.MongoTenantPredicateInjector
+dev.caskeleton.adapter.outbound.mongo.advanced.tenancy.shared.TenantScopedMongoOperations
+dev.caskeleton.adapter.outbound.mongo.advanced.timeseries.MongoTimeSeriesCapability
+dev.caskeleton.adapter.outbound.mongo.advanced.timeseries.MongoTimeSeriesCapabilityValidator
+dev.caskeleton.adapter.outbound.mongo.advanced.timeseries.MongoTimeSeriesDescriptor
+dev.caskeleton.adapter.outbound.mongo.advanced.timeseries.MongoTimeSeriesGranularity
+dev.caskeleton.adapter.outbound.mongo.advanced.timeseries.MongoTimeSeriesOperations
+dev.caskeleton.adapter.outbound.mongo.advanced.timeseries.MongoTimeSeriesSupport
+dev.caskeleton.adapter.outbound.mongo.advanced.vector.MongoEmbedding
+dev.caskeleton.adapter.outbound.mongo.advanced.vector.MongoVectorIndexDescriptor
+dev.caskeleton.adapter.outbound.mongo.advanced.vector.MongoVectorQuery
+dev.caskeleton.adapter.outbound.mongo.advanced.vector.MongoVectorSearchBenchmarkGate
+dev.caskeleton.adapter.outbound.mongo.advanced.vector.MongoVectorSearchOperations
+dev.caskeleton.adapter.outbound.mongo.aggregation.MongoAggregationPlan
+dev.caskeleton.adapter.outbound.mongo.aggregation.MongoAggregationProfile
+dev.caskeleton.adapter.outbound.mongo.aggregation.MongoAggregationRisk
+dev.caskeleton.adapter.outbound.mongo.aggregation.MongoAggregationStageDescriptor
+dev.caskeleton.adapter.outbound.mongo.aggregation.PolicyAwareMongoAggregationExecutor
+dev.caskeleton.adapter.outbound.mongo.api.CollectionProfileName
+dev.caskeleton.adapter.outbound.mongo.api.DatabaseProfileName
+dev.caskeleton.adapter.outbound.mongo.api.MongoOperationContext
+dev.caskeleton.adapter.outbound.mongo.api.MongoOperationName
+dev.caskeleton.adapter.outbound.mongo.api.MongoOperationScope
+dev.caskeleton.adapter.outbound.mongo.api.MongoOperationType
+dev.caskeleton.adapter.outbound.mongo.api.capability.MongoCapability
+dev.caskeleton.adapter.outbound.mongo.api.capability.MongoCapabilitySet
+dev.caskeleton.adapter.outbound.mongo.api.capability.MongoCapabilitySupport
+dev.caskeleton.adapter.outbound.mongo.api.capability.MongoServerVersion
+dev.caskeleton.adapter.outbound.mongo.api.capability.MongoSupportLevel
+dev.caskeleton.adapter.outbound.mongo.api.consistency.MongoConsistencyDescriptor
+dev.caskeleton.adapter.outbound.mongo.api.consistency.MongoConsistencyGuarantee
+dev.caskeleton.adapter.outbound.mongo.api.consistency.MongoConsistencyProfile
+dev.caskeleton.adapter.outbound.mongo.api.consistency.MongoConsistencyRegistry
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoBulkPartialFailureException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoConnectionException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoCursorException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoDataSchemaUnsupportedException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoDocumentTooLargeException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoDuplicateKeyException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoEncryptionException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoExecutionOutcome
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoFailureCategory
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoFailureContext
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoOperationRejectedException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoOptimisticConflictException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoPersistenceException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoReadConcernException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoResumeException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoRetryScope
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoSchemaValidationException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoServerSelectionException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoShardRoutingException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoTimeoutException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoTransactionCommitUnknownException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoTransactionTransientException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoUnclassifiedFailureException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoWriteConcernException
+dev.caskeleton.adapter.outbound.mongo.api.error.MongoWriteConflictException
+dev.caskeleton.adapter.outbound.mongo.api.mapping.DomainDocumentId
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoBigIntegerRepresentation
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoDecimalRepresentation
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoEnumRepresentation
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoIdRepresentation
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoTemporalRepresentation
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoTypeMetadataPolicy
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoTypeRepresentationManifest
+dev.caskeleton.adapter.outbound.mongo.api.mapping.MongoUuidRepresentation
+dev.caskeleton.adapter.outbound.mongo.api.observation.MongoOperationObservation
+dev.caskeleton.adapter.outbound.mongo.api.observation.MongoOperationObserver
+dev.caskeleton.adapter.outbound.mongo.api.profile.MongoClientPlane
+dev.caskeleton.adapter.outbound.mongo.api.profile.MongoRuntimeProfile
+dev.caskeleton.adapter.outbound.mongo.api.profile.MongoStableApiProfile
+dev.caskeleton.adapter.outbound.mongo.api.profile.MongoTopology
+dev.caskeleton.adapter.outbound.mongo.api.profile.MongoTopologyRequirement
+dev.caskeleton.adapter.outbound.mongo.api.schema.DocumentSchemaVersion
+dev.caskeleton.adapter.outbound.mongo.api.schema.MongoSchemaVersionPolicy
+dev.caskeleton.adapter.outbound.mongo.api.schema.MongoSchemaVersionRange
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoClientGeneration
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoClientGenerationRegistry
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoDriverObservabilityAutoConfiguration
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoPlatformAutoConfiguration
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoPlatformHealthIndicator
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoPlatformSettings
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoProfileProperties
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoStartupValidator
+dev.caskeleton.adapter.outbound.mongo.autoconfigure.MongoTopologyProbe
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoChangeEventIdentity
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoChangeStreamPipeline
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoChangeStreamState
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoChangeStreamSubscription
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoClusterTime
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoResumeCheckpoint
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoResumeCheckpointStore
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoResumePosition
+dev.caskeleton.adapter.outbound.mongo.changestream.MongoResumeTokenCodec
+dev.caskeleton.adapter.outbound.mongo.changestream.consumer.MongoChangeStreamSource
+dev.caskeleton.adapter.outbound.mongo.changestream.consumer.ReactiveMongoChangeStreamConsumer
+dev.caskeleton.adapter.outbound.mongo.changestream.consumer.SpringReactiveChangeStreamSource
+dev.caskeleton.adapter.outbound.mongo.changestream.projector.MongoChangeClaim
+dev.caskeleton.adapter.outbound.mongo.changestream.projector.MongoChangeDeduplicationStore
+dev.caskeleton.adapter.outbound.mongo.changestream.projector.MongoChangeProjectionResult
+dev.caskeleton.adapter.outbound.mongo.changestream.projector.MongoChangeProjector
+dev.caskeleton.adapter.outbound.mongo.changestream.projector.MongoChangeStreamRunner
+dev.caskeleton.adapter.outbound.mongo.changestream.recovery.MongoChangeHistoryLostException
+dev.caskeleton.adapter.outbound.mongo.changestream.recovery.MongoChangeStreamRecoveryDecision
+dev.caskeleton.adapter.outbound.mongo.changestream.recovery.MongoChangeStreamRecoveryPolicy
+dev.caskeleton.adapter.outbound.mongo.changestream.recovery.MongoInvalidateRecovery
+dev.caskeleton.adapter.outbound.mongo.client.MongoClientSettingsFactory
+dev.caskeleton.adapter.outbound.mongo.failure.DefaultMongoFailureClassifier
+dev.caskeleton.adapter.outbound.mongo.failure.DefaultMongoFailureTranslator
+dev.caskeleton.adapter.outbound.mongo.failure.MongoDriverFailureView
+dev.caskeleton.adapter.outbound.mongo.failure.MongoFailureClassification
+dev.caskeleton.adapter.outbound.mongo.failure.MongoFailureClassifier
+dev.caskeleton.adapter.outbound.mongo.failure.MongoFailureExtractor
+dev.caskeleton.adapter.outbound.mongo.failure.MongoFailurePhase
+dev.caskeleton.adapter.outbound.mongo.failure.MongoFailureTranslator
+dev.caskeleton.adapter.outbound.mongo.geo.MongoGeoDistance
+dev.caskeleton.adapter.outbound.mongo.geo.MongoGeoPoint
+dev.caskeleton.adapter.outbound.mongo.geo.MongoGeoQuery
+dev.caskeleton.adapter.outbound.mongo.geo.MongoGeospatialOperations
+dev.caskeleton.adapter.outbound.mongo.geo.SpringMongoGeospatialOperations
+dev.caskeleton.adapter.outbound.mongo.imperative.DefaultMongoImperativeExecutor
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoCollectionAccess
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoCollectionProfileRegistry
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoCompletion
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoConsistencyBinder
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoImperativeCallback
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoImperativeExecutor
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoOperationResult
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoPlatformCallback
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoPlatformCollectionAccess
+dev.caskeleton.adapter.outbound.mongo.imperative.MongoTemplateSupportContract
+dev.caskeleton.adapter.outbound.mongo.imperative.ScopedMongoOperations
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.AtomicFilter
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.AtomicUpdate
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.AtomicUpdateResult
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.MongoAtomicOperations
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.MongoAtomicOperationsTemplate
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.MongoAtomicPolicy
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.MongoAtomicPolicyRegistry
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.MongoUpdateOperator
+dev.caskeleton.adapter.outbound.mongo.imperative.atomic.ReturnDocumentMode
+dev.caskeleton.adapter.outbound.mongo.imperative.bulk.MongoBulkExecutor
+dev.caskeleton.adapter.outbound.mongo.imperative.bulk.MongoBulkItemFailure
+dev.caskeleton.adapter.outbound.mongo.imperative.bulk.MongoBulkItemOutcome
+dev.caskeleton.adapter.outbound.mongo.imperative.bulk.MongoBulkMode
+dev.caskeleton.adapter.outbound.mongo.imperative.bulk.MongoBulkResult
+dev.caskeleton.adapter.outbound.mongo.imperative.bulk.MongoBulkWritePlan
+dev.caskeleton.adapter.outbound.mongo.imperative.bulk.SpringDataBulkFailureExtractor
+dev.caskeleton.adapter.outbound.mongo.imperative.revision.MongoDocumentNotFoundException
+dev.caskeleton.adapter.outbound.mongo.imperative.revision.MongoOptimisticConflictTranslator
+dev.caskeleton.adapter.outbound.mongo.imperative.revision.MongoRevision
+dev.caskeleton.adapter.outbound.mongo.imperative.revision.VersionedMongoUpdater
+dev.caskeleton.adapter.outbound.mongo.imperative.revision.VersionedUpdateCommand
+dev.caskeleton.adapter.outbound.mongo.mapping.BigDecimalToDecimal128Converter
+dev.caskeleton.adapter.outbound.mongo.mapping.BigIntegerRepresentationConverters
+dev.caskeleton.adapter.outbound.mongo.mapping.Decimal128ToBigDecimalConverter
+dev.caskeleton.adapter.outbound.mongo.mapping.DomainIdReadConverter
+dev.caskeleton.adapter.outbound.mongo.mapping.DomainIdWriteConverter
+dev.caskeleton.adapter.outbound.mongo.mapping.LocalDateTimeMappingGuard
+dev.caskeleton.adapter.outbound.mongo.mapping.MongoCustomConversionsFactory
+dev.caskeleton.adapter.outbound.mongo.mapping.MongoMappingConfiguration
+dev.caskeleton.adapter.outbound.mongo.mapping.MongoTypeMetadataConfigurer
+dev.caskeleton.adapter.outbound.mongo.mapping.type.LongLivedMongoDocument
+dev.caskeleton.adapter.outbound.mongo.mapping.type.MongoTypeMetadataDescriptor
+dev.caskeleton.adapter.outbound.mongo.mapping.type.MongoTypeMetadataRegistry
+dev.caskeleton.adapter.outbound.mongo.mapping.type.PolicyAwareMongoTypeMapper
+dev.caskeleton.adapter.outbound.mongo.migration.MongoCollectionMigrationLedger
+dev.caskeleton.adapter.outbound.mongo.migration.MongoCollectionMigrationLock
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigration
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationCheckpoint
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationChecksum
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationContext
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationHeartbeat
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationId
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationLedger
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationLock
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationPostcondition
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationPrecondition
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationResult
+dev.caskeleton.adapter.outbound.mongo.migration.MongoMigrationRunner
+dev.caskeleton.adapter.outbound.mongo.migration.flamingock.FlamingockChangeUnitView
+dev.caskeleton.adapter.outbound.mongo.migration.flamingock.FlamingockLedgerAdapter
+dev.caskeleton.adapter.outbound.mongo.migration.flamingock.FlamingockLockAdapter
+dev.caskeleton.adapter.outbound.mongo.migration.flamingock.FlamingockMigrationConfiguration
+dev.caskeleton.adapter.outbound.mongo.migration.flamingock.FlamingockMongoMigrationAdapter
+dev.caskeleton.adapter.outbound.mongo.nativecap.ApprovedMongoNativeOperation
+dev.caskeleton.adapter.outbound.mongo.nativecap.MongoNativeCapabilityGateway
+dev.caskeleton.adapter.outbound.mongo.nativecap.MongoNativeCommandCategory
+dev.caskeleton.adapter.outbound.mongo.nativecap.MongoNativeOperationPolicy
+dev.caskeleton.adapter.outbound.mongo.nativecap.PolicyAwareMongoNativeGateway
+dev.caskeleton.adapter.outbound.mongo.observation.MicrometerMongoOperationObserver
+dev.caskeleton.adapter.outbound.mongo.observation.MongoCommandObservationListener
+dev.caskeleton.adapter.outbound.mongo.observation.MongoDriverObservabilityConfiguration
+dev.caskeleton.adapter.outbound.mongo.observation.MongoObservationConvention
+dev.caskeleton.adapter.outbound.mongo.observation.MongoObservationRedactor
+dev.caskeleton.adapter.outbound.mongo.observation.MongoPoolObservationListener
+dev.caskeleton.adapter.outbound.mongo.observation.MongoSdamObservationListener
+dev.caskeleton.adapter.outbound.mongo.query.MongoFieldDescriptor
+dev.caskeleton.adapter.outbound.mongo.query.MongoOperator
+dev.caskeleton.adapter.outbound.mongo.query.MongoQueryPolicy
+dev.caskeleton.adapter.outbound.mongo.query.MongoRegexPolicy
+dev.caskeleton.adapter.outbound.mongo.query.MongoSortDescriptor
+dev.caskeleton.adapter.outbound.mongo.query.PolicyAwareMongoQueryBuilder
+dev.caskeleton.adapter.outbound.mongo.query.budget.MongoBudgetEnforcer
+dev.caskeleton.adapter.outbound.mongo.query.budget.MongoBudgetPolicyRegistry
+dev.caskeleton.adapter.outbound.mongo.query.budget.MongoOperationBudget
+dev.caskeleton.adapter.outbound.mongo.query.pagination.MongoKeysetCursor
+dev.caskeleton.adapter.outbound.mongo.query.pagination.MongoKeysetCursorCodec
+dev.caskeleton.adapter.outbound.mongo.query.pagination.MongoKeysetPageRequest
+dev.caskeleton.adapter.outbound.mongo.query.pagination.MongoKeysetQueryBuilder
+dev.caskeleton.adapter.outbound.mongo.query.pagination.MongoKeysetSlice
+dev.caskeleton.adapter.outbound.mongo.query.pagination.MongoKeysetSort
+dev.caskeleton.adapter.outbound.mongo.query.pagination.MongoNullSortOrdering
+dev.caskeleton.adapter.outbound.mongo.reactive.DefaultReactiveMongoExecutor
+dev.caskeleton.adapter.outbound.mongo.reactive.ReactiveMongoCallback
+dev.caskeleton.adapter.outbound.mongo.reactive.ReactiveMongoCollectionAccess
+dev.caskeleton.adapter.outbound.mongo.reactive.ReactiveMongoConsistencyBinder
+dev.caskeleton.adapter.outbound.mongo.reactive.ReactiveMongoContextKeys
+dev.caskeleton.adapter.outbound.mongo.reactive.ReactiveMongoExecutor
+dev.caskeleton.adapter.outbound.mongo.reactive.ReactiveScopedMongoOperations
+dev.caskeleton.adapter.outbound.mongo.reactive.cursor.MongoCursorGuard
+dev.caskeleton.adapter.outbound.mongo.reactive.cursor.MongoCursorLease
+dev.caskeleton.adapter.outbound.mongo.reactive.cursor.MongoCursorTermination
+dev.caskeleton.adapter.outbound.mongo.reactive.cursor.MongoReactiveCursorPublisher
+dev.caskeleton.adapter.outbound.mongo.reactive.cursor.MongoResultBudgetTracker
+dev.caskeleton.adapter.outbound.mongo.schema.index.MongoIndexApplyPolicy
+dev.caskeleton.adapter.outbound.mongo.schema.index.MongoIndexDescriptorView
+dev.caskeleton.adapter.outbound.mongo.schema.index.MongoIndexDiff
+dev.caskeleton.adapter.outbound.mongo.schema.index.MongoIndexDiffEngine
+dev.caskeleton.adapter.outbound.mongo.schema.index.MongoIndexRetirementPlan
+dev.caskeleton.adapter.outbound.mongo.schema.index.MongoIndexRetirementState
+dev.caskeleton.adapter.outbound.mongo.schema.manifest.MongoCollectionManifest
+dev.caskeleton.adapter.outbound.mongo.schema.manifest.MongoIndexDirection
+dev.caskeleton.adapter.outbound.mongo.schema.manifest.MongoIndexKey
+dev.caskeleton.adapter.outbound.mongo.schema.manifest.MongoIndexManifest
+dev.caskeleton.adapter.outbound.mongo.schema.manifest.MongoManifestRegistry
+dev.caskeleton.adapter.outbound.mongo.schema.manifest.MongoMetadataOwnership
+dev.caskeleton.adapter.outbound.mongo.schema.manifest.MongoSchemaManifest
+dev.caskeleton.adapter.outbound.mongo.schema.model.EmbeddedCollectionDescriptor
+dev.caskeleton.adapter.outbound.mongo.schema.model.MongoBinaryFieldDescriptor
+dev.caskeleton.adapter.outbound.mongo.schema.model.MongoDocumentModelManifest
+dev.caskeleton.adapter.outbound.mongo.schema.model.MongoDocumentModelValidator
+dev.caskeleton.adapter.outbound.mongo.schema.model.MongoDocumentSizeBudget
+dev.caskeleton.adapter.outbound.mongo.schema.model.MongoReferenceDescriptor
+dev.caskeleton.adapter.outbound.mongo.schema.model.MongoReferenceLifecycle
+dev.caskeleton.adapter.outbound.mongo.schema.ttl.MongoExpirationAccessPolicy
+dev.caskeleton.adapter.outbound.mongo.schema.ttl.MongoTtlIndexDescriptor
+dev.caskeleton.adapter.outbound.mongo.schema.ttl.MongoTtlPolicy
+dev.caskeleton.adapter.outbound.mongo.schema.ttl.MongoTtlPolicyValidator
+dev.caskeleton.adapter.outbound.mongo.schema.validation.MongoValidationAction
+dev.caskeleton.adapter.outbound.mongo.schema.validation.MongoValidationLevel
+dev.caskeleton.adapter.outbound.mongo.schema.validation.MongoValidatorApplyPolicy
+dev.caskeleton.adapter.outbound.mongo.schema.validation.MongoValidatorDescriptor
+dev.caskeleton.adapter.outbound.mongo.schema.validation.MongoValidatorDiff
+dev.caskeleton.adapter.outbound.mongo.schema.validation.MongoValidatorDiffEngine
+dev.caskeleton.adapter.outbound.mongo.security.MongoCredentialReference
+dev.caskeleton.adapter.outbound.mongo.security.MongoCredentialResolver
+dev.caskeleton.adapter.outbound.mongo.security.MongoCredentialRotationPolicy
+dev.caskeleton.adapter.outbound.mongo.security.MongoPrincipalRole
+dev.caskeleton.adapter.outbound.mongo.security.MongoSecurityProfile
+dev.caskeleton.adapter.outbound.mongo.security.MongoSecurityProfileValidator
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminApproval
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminAuditPhase
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminAuditRecord
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminAuthorization
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminCommand
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminGateway
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminOperation
+dev.caskeleton.adapter.outbound.mongo.security.admin.MongoAdminRuntimeGuard
+dev.caskeleton.adapter.outbound.mongo.transaction.MongoTransactionExecutor
+dev.caskeleton.adapter.outbound.mongo.transaction.MongoTransactionProfile
+dev.caskeleton.adapter.outbound.mongo.transaction.MongoTransactionScope
+dev.caskeleton.adapter.outbound.mongo.transaction.MongoTransactionSession
+dev.caskeleton.adapter.outbound.mongo.transaction.MongoTransactionSessionFactory
+dev.caskeleton.adapter.outbound.mongo.transaction.ReactiveMongoTransactionExecutor
+dev.caskeleton.adapter.outbound.mongo.transaction.ReactiveMongoTransactionSession
+dev.caskeleton.adapter.outbound.mongo.transaction.ReactiveMongoTransactionSessionFactory
+dev.caskeleton.adapter.outbound.mongo.transaction.SpringMongoTransactionExecutor
+dev.caskeleton.adapter.outbound.mongo.transaction.SpringMongoTransactionSessionFactory
+dev.caskeleton.adapter.outbound.mongo.transaction.SpringReactiveMongoTransactionExecutor
+dev.caskeleton.adapter.outbound.mongo.transaction.SpringReactiveMongoTransactionSessionFactory
+dev.caskeleton.adapter.outbound.mongo.transaction.retry.MongoCommitReconciler
+dev.caskeleton.adapter.outbound.mongo.transaction.retry.MongoRetryBudget
+dev.caskeleton.adapter.outbound.mongo.transaction.retry.MongoRetryDecision
+dev.caskeleton.adapter.outbound.mongo.transaction.retry.MongoTransactionRetryCoordinator
+dev.caskeleton.adapter.outbound.mongo.transaction.session.MongoCausalSessionContext
+dev.caskeleton.adapter.outbound.mongo.transaction.session.MongoCausalSessionExecutor
+dev.caskeleton.adapter.outbound.mongo.transaction.session.ReactiveMongoCausalSessionExecutor
+dev.caskeleton.adapter.outbound.mongo.transaction.session.SpringMongoCausalSessionExecutor
@@ -0,0 +1,73 @@
+# CI/CD 경계 — GitHub Actions는 CI, Argo CD는 CD
+
+## 결론
+
+GitHub Actions는 **검증하고 아티팩트를 만든다**. Argo CD는 **배포한다**. 두 역할은 겹치지 않는다.
+
+GitHub Actions 워크플로는 `kubectl apply`, `helm upgrade`, `argocd app sync` 중 어느 것도 하지
+않는다. 그러므로 CI에는 클러스터 자격증명(kubeconfig, 서비스 계정 토큰)이 들어가지 않는다.
+
+## 흐름
+
+```text
+   git push / tag
+        │
+        ▼
+  GitHub Actions  ─────────────── CI ───────────────┐
+   • 테스트 · 정적분석 · 아키텍처 검증               │
+   • 컨테이너 이미지 빌드                            │
+   • 취약점 스캔 (Trivy)                             │
+   • SBOM 생성                                       │
+   • 레지스트리에 이미지 push                        │
+        │                                            │
+        │ 이미지 태그(다이제스트)를 manifest에 기록  │
+        ▼                                            │
+  GitOps 저장소 (배포 희망 상태)  ──────────────────┘
+        │
+        │ Argo CD가 watch
+        ▼
+     Argo CD  ──────────────────── CD ───────────────
+        │ auto-sync
+        ▼
+   Kubernetes
+```
+
+용어 한 줄 풀이:
+
+- **GitOps 저장소** — 클러스터에 무엇이 떠 있어야 하는지를 적어 둔 Git 저장소. 애플리케이션 소스와
+  분리한다.
+- **manifest** — Kubernetes에 넣을 YAML(Deployment, Service 등).
+- **auto-sync** — Argo CD가 GitOps 저장소의 변경을 스스로 감지해 클러스터에 반영하는 모드. 이걸 쓰면
+  CI가 Argo CD API 서버에 접근할 필요가 없다.
+
+## 왜 이렇게 나누나
+
+1. **자격증명 반경.** CI가 배포하면 CI 러너가 프로덕션 클러스터에 대한 쓰기 권한을 갖는다. 포크된
+   PR, 서드파티 액션, 캐시 오염이 모두 그 권한에 닿는다. auto-sync를 쓰면 그 권한은 클러스터 안의
+   Argo CD에만 있고, CI는 Git에 커밋만 한다.
+2. **현재 상태의 소유자가 하나.** 클러스터에 무엇이 떠 있는지는 GitOps 저장소가 답한다. CI가 직접
+   apply 하면 답이 두 개가 된다 — Git에 적힌 것과 실제로 떠 있는 것.
+3. **롤백이 revert.** 배포를 되돌리는 것이 `git revert`가 된다.
+
+## 이 저장소의 현재 위치
+
+| 항목 | 상태 |
+| --- | --- |
+| 이미지 빌드/스캔/push | `release.yml`이 수행 |
+| SBOM | `release.yml`이 생성 |
+| 이미지 서명 · provenance attestation | **없음.** 추가 대상 |
+| GitOps 저장소 | **없음.** 별도 저장소로 만들 예정 |
+| Argo CD Application 정의 | **없음.** GitOps 저장소에 둘 예정 |
+| CI에서의 클러스터 접근 | 없음 — 유일했던 `kubectl apply`는 제거됨 |
+
+`fileserver-certification.yml`은 예외처럼 보이지만 아니다. PVC 매니페스트가 여전히 ReadWriteOnce를
+선언하는지 **파일만** 확인하고, 클러스터에는 아무것도 적용하지 않는다. 실제 클러스터에서의 인증은
+운영자가 `infra/fileserver/kubernetes/pvc-certification-job.yaml`을 직접 실행하고
+`docs/fileserver/storage-certification.md`에 기록한다. 이름을 `fileserver-release.yml`에서 바꾼 이유가
+이것이다 — 이 워크플로는 릴리스하지 않는다.
+
+## 규칙
+
+- 워크플로에 클러스터 자격증명 secret을 추가하지 않는다.
+- 배포 대상이 바뀌면 GitOps 저장소의 manifest를 바꾼다. 워크플로를 바꾸지 않는다.
+- CI가 만드는 것은 **불변 다이제스트로 지정된 이미지**다. `latest` 태그로 배포하지 않는다.
@@ -0,0 +1,50 @@
+# Template maintainer와 Template consumer의 검증은 다르다
+
+## 결론
+
+이 저장소에는 성격이 다른 두 종류의 검증이 섞여 있다.
+
+1. **스켈레톤을 만드는 사람**에게 필요한 검증 — sample 모듈이 정말 제거 가능한가, optional 모듈
+   조합이 모두 빌드되는가, 레지스트리가 확장 가능한가.
+2. **스켈레톤을 가져다 서비스를 만드는 사람**에게 필요한 검증 — 내 애플리케이션의 테스트,
+   아키텍처 방향, 보안, 릴리스.
+
+파생 프로젝트가 1번을 그대로 물려받으면, 자기 서비스와 아무 상관 없는 게이트를 평생 유지하게 된다.
+이 문서는 어느 쪽이 어느 쪽인지 적어 둔다.
+
+## Template 전용 (파생 프로젝트는 삭제해도 된다)
+
+| 대상 | 무엇을 지키는가 |
+| --- | --- |
+| `:app-bootstrap:sampleOffTest`, `ci-quality-gates.yml`의 `sample-off` job | sample 픽스처를 지워도 애플리케이션이 빌드·부팅되는가 |
+| `sample-portfolio` leaf 전체 | 참조 구현 |
+| `Dockerfile.sample`, `docker-compose.*` 중 sample 관련 | 위와 동일 |
+| `docs/superpowers/**` | 이 템플릿을 만든 과정의 설계/계획 기록 |
+| `gradle/qualification/**` | 이 템플릿이 벤더링한 플랫폼(JPA, messaging)의 인증 체계 |
+| `*-certification.yml`, `*-qualification.yml`, `jpa-next-*.yml` | 템플릿이 광고하는 지원 매트릭스의 근거 |
+
+## Consumer 필수 (파생 프로젝트가 유지해야 한다)
+
+| 대상 | 무엇을 지키는가 |
+| --- | --- |
+| `architectureCheck` | Clean Architecture 의존 방향. 이 템플릿의 존재 이유 |
+| 각 leaf의 `check` | 컴파일 · 단위 테스트 · 포맷 · 스타일 · Error Prone |
+| `qualityCheck` | SpotBugs / FindSecBugs |
+| `configContractCheck` | 환경변수 계약 |
+| `verifyDependencyLocks` | 재현 가능한 의존성 해석 |
+| `dependency-vulnerability.yml` | dependency-review + Trivy |
+| `ci-quality-gates.yml` | PR 게이트 |
+| `release.yml` | 이미지 · SBOM 생산 |
+| action의 full SHA 핀 | 공급망 |
+
+## 파생 프로젝트가 할 일
+
+1. Template 전용 표의 항목을 삭제한다. 삭제는 대부분 파일 삭제 + `config/architecture/modules.json`
+   에서 leaf 항목 제거로 끝난다 — 레지스트리가 leaf 목록의 SSOT이고, 개수를 따로 적어 둔 곳은 없다.
+2. `docs/ci-cd/boundary.md`의 경계를 그대로 유지한 채 자기 GitOps 저장소를 연결한다.
+3. `.trivyignore.yaml`과 CODEOWNERS는 그대로 쓴다.
+
+## 아직 하지 않은 것
+
+Template CI와 Generated Application CI를 **물리적으로** 분리하지는 않았다(생성기 없음). 지금은 이
+문서가 그 경계다. 생성기를 만든다면, 위 표의 "Template 전용" 열이 생성기가 벗겨 내야 할 목록이다.
@@ -0,0 +1,62 @@
+# gRPC advanced capability support matrix
+
+Every capability in `:grpc-advanced:*`, its grade, and what it would take to raise it.
+`GrpcAdvancedSupportMatrix` is the machine-readable form; `GrpcAdvancedCapability.defaultGrade`
+carries the same values.
+
+All capabilities are off by default. Flags are `ca-skeleton.grpc.advanced.<capability>.enabled`.
+
+## Grades
+
+| Grade | May start | Production needs a separate approval |
+| --- | --- | --- |
+| `ADVANCED_STABLE` | Yes | No |
+| `EXPERIMENTAL` | Yes | Yes |
+| `WATCH` | No | — |
+| `DISABLED` | No | — |
+
+## Capabilities
+
+| Capability | Flag | Grade | Real infrastructure its evidence needs |
+| --- | --- | --- | --- |
+| Protobuf Edition 2024 | `edition-2024` | `ADVANCED_STABLE` | — |
+| Protobuf Edition 2026 | `edition-2026` | `WATCH` | — |
+| Client streaming | `client-streaming` | `ADVANCED_STABLE` | — |
+| Bidirectional streaming | `bidi-streaming` | `ADVANCED_STABLE` | — |
+| Manual flow control | `manual-flow-control` | `ADVANCED_STABLE` | — |
+| Read-only unary hedging | `hedging` | `EXPERIMENTAL` | — |
+| Custom name resolver | `custom-resolver` | `ADVANCED_STABLE` | — |
+| Custom load balancer | `custom-load-balancer` | `EXPERIMENTAL` | — |
+| Proxyless xDS | `xds` | `EXPERIMENTAL` | xDS control plane |
+| gRPC-Web | `grpc-web` | `ADVANCED_STABLE` | gRPC-Web proxy |
+| Servlet HTTP/2 | `servlet-compat` | `ADVANCED_STABLE` | Servlet container |
+| Spring Integration bridge | `integration-bridge` | `ADVANCED_STABLE` | — |
+| Reactor adapter | `reactor` | `ADVANCED_STABLE` | — |
+| Kotlin coroutine / Flow | `kotlin` | `ADVANCED_STABLE` | Kotlin toolchain |
+| Channelz / CSDS diagnostics | `channel-diagnostics` | `ADVANCED_STABLE` | — |
+
+## What the grades mean here, concretely
+
+**Grade is a statement about the contract, not about a deployment.** Every capability's contract is
+implemented and tested in this repository. What no capability has is evidence from a real deployment:
+`GrpcAdvancedPromotionEvidence` for each one is empty, and no promotion has been granted.
+
+**Four capabilities cannot produce meaningful evidence here at all**, because the infrastructure they
+need is absent. `GrpcAdvancedInfrastructureTestkit.missingInfrastructure` names them, and a suite that
+runs without its infrastructure passes and establishes nothing.
+
+**Kotlin is the sharpest case.** This repository has no Kotlin toolchain, so
+`GrpcKotlinCompatibilityGate.supportableHere()` returns false and always will until one exists. The
+four contract requirements — one schema source shared with Java, coroutine cancellation propagated,
+Flow backpressure inside the Stable buffer bounds, platform evidence types preserved — are checkable
+without a toolchain and are checked. The compile lane is not.
+
+## Promotion thresholds
+
+| To | Soak | Also required |
+| --- | --- | --- |
+| `ADVANCED_STABLE` | 7 days | compatibility evidence, security review, fault evidence, performance evidence, ADR, runbook, real-environment test |
+| Stable default | 30 days | all of the above, plus a dependency, security and operational-cost review |
+
+`WATCH` becomes `EXPERIMENTAL` before anything else. Promotions are independent: promoting one
+capability changes no other's grade.
@@ -0,0 +1,75 @@
+# gRPC platform support matrix
+
+What the Stable gRPC platform (`:grpc:*`) is certified against, what it is only checked against, and
+what is merely watched. The distinction is the point: "works with Spring Boot" is not a statement
+anyone can act on.
+
+`GrpcCompatibilityMatrix.caSkeleton()` is the machine-readable form of this table, and
+`GrpcStableReleaseGate` blocks a release when a certified lane has no result or a failing one.
+
+## Lanes
+
+| Lane | Grade | Failure blocks a release |
+| --- | --- | --- |
+| Boot-managed platform (Spring Boot 4.0.8 BOM) | Certified | Yes |
+| proto3 with explicit `optional` | Certified | Yes |
+| `grpc-netty-shaded` | Certified | Yes |
+| `grpc-netty` (unshaded) | Compatibility | No |
+| Upstream gRPC Java version override | Compatibility | No |
+| Protobuf Edition 2024 | Watch | No |
+| Protobuf Edition 2026 | Watch | No |
+
+## Runtime baseline
+
+| | |
+| --- | --- |
+| Java | 21 |
+| Spring Boot | 4.0.8 (the repository baseline; the plans assume 4.1) |
+| io.grpc | `ext.grpcVersion` in `src/build.gradle` |
+| Protobuf | `ext.protobufVersion` in `src/build.gradle` |
+| Stable transport | Netty (shaded) |
+| Stable RPC shapes | Unary, Server Streaming |
+| Stable resolvers | Static, DNS, Unix domain socket |
+| Stable load balancing | `pick_first`, `round_robin` |
+
+## Evidence grades
+
+A capability may only be advertised on evidence of a grade that can establish it.
+`GrpcEvidenceGrade.requireCertifies` enforces this, and `GrpcReleaseEvidence.supports` refuses a
+claim backed by the wrong lane.
+
+| Grade | Lane | Establishes |
+| --- | --- | --- |
+| `CONTRACT` | `grpcInProcessContractTest` | adapter, interceptor order, status mapping, validation, idempotency replay, context propagation |
+| `TRANSPORT` | `grpcNettyContractTest` | HTTP/2, TLS, mTLS, metadata limit, message limit, GOAWAY, keepalive, graceful shutdown |
+| `FAULT` | `grpcFaultTest` | connection loss, completion unknown, partial stream, evidence classifier |
+| `PERFORMANCE` | `grpcPerformanceTest` | latency, stream saturation, executor saturation, drain budget |
+
+In-process results are never transport evidence. The in-process transport does not negotiate TLS,
+does not frame HTTP/2 and does not enforce transport-level limits, so a suite that passes there has
+tested the adapter and not the transport.
+
+## What is not supported
+
+| | Where it lives |
+| --- | --- |
+| Client streaming, bidirectional streaming | `grpc-advanced-streaming` |
+| Manual flow control | `grpc-advanced-streaming` |
+| Hedging | `grpc-advanced-resilience` |
+| Custom name resolver, custom load balancer | `grpc-advanced-resilience` |
+| xDS | `grpc-advanced-resilience` |
+| gRPC-Web, Servlet HTTP/2, Spring Integration, Reactor, Kotlin | `grpc-advanced-compat` |
+| Channelz / CSDS diagnostics | `grpc-advanced-diagnostics` |
+
+## Current release status
+
+Not released. Every `:grpc:*` leaf is `runtime_memberships: []` in the module registry, so the
+platform is build-only: it compiles, its lanes run, and no deployed artifact carries it.
+
+Two release gate inputs are outstanding and are the work between here and a release:
+
+- **Performance baseline.** The performance lane runs and asserts shape — ordered percentiles, a gate
+  that reads them — rather than absolute numbers. A recorded baseline on a known runner is what turns
+  it into a regression gate.
+- **Schema codegen.** No `protoc` runs in this build (ADR-GRPC-002), so the descriptor artifact and
+  the consumer-compile fixture are governed as policy rather than produced from a compiled schema.
@@ -0,0 +1,176 @@
+# Fileserver configuration
+
+Every key below lives under `app.fileserver-platform` (environment form
+`APP_FILESERVER_PLATFORM_*`). That namespace is the HTTP platform's alone: `app.fileserver.*`
+belongs to the R2 tabular publication capability and `app.file-export.*` to the R1 CSV export, and
+the three are deliberately separate so switching one on cannot switch on another.
+
+While `app.fileserver-platform.enabled` is false none of these keys is bound at all — the
+auto-configuration that binds them is not processed — so a malformed value in a block nobody
+enabled cannot fail a startup. Once enabled, binding is strict: an unknown key under the prefix is
+refused rather than ignored. The defaults are the conservative ones: the
+capability is off, the admin plane is off, background reclamation is off, and there is no permissive
+authorization fallback. Turning the capability on is a deliberate act, and so is every surface it
+exposes.
+
+## Minimum to start
+
+```yaml
+ca-skeleton:
+  fileserver:
+    enabled: true
+    instance-id: ${HOSTNAME}                       # writer-lease owner; must be unique per node
+    storage:
+      root: /var/lib/backend/files                 # absolute, outside any webroot or config dir
+    security:
+      access-policy: role-based                    # or supply your own FileAccessPolicy bean
+    observability:
+      fingerprint-key: ${FILESERVER_FINGERPRINT_KEY}
+```
+
+Startup fails, rather than degrading, when any of these is missing or unsafe:
+
+| Condition | Why it is fatal |
+| --- | --- |
+| `security.access-policy` left at `required` with no `FileAccessPolicy` bean | a file capability that authorizes by default is worse than one that refuses to start |
+| `observability.fingerprint-key` unset while metrics are on | an unkeyed digest of an enumerable identifier is reversible |
+| the storage root fails a mandatory capability probe | a volume that cannot create atomically, keep staging and content on one FileStore, or refuse symlinks is unsafe, not degraded |
+| `storage.publish-mode: atomic-move-required` on a volume where the probe could not prove an atomic move | the configured guarantee cannot be delivered |
+| `security.access-policy: unenforced` under a `prod` profile | a value that was convenient in development must not survive promotion |
+
+## Authorization — `security`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `access-policy` | `required` | `required` (supply your own bean), `role-based`, or `unenforced` |
+| `read-roles` | `ROLE_FILE_READ` | grants metadata read and download |
+| `write-roles` | `ROLE_FILE_WRITE` | grants create, append, finalize, delete, copy, move |
+| `admin-roles` | `ROLE_FILE_ADMIN` | grants reverify and force-delete, and gates `/internal/fileserver/**` at the servlet chain |
+
+There is no anonymous-read switch. Every Fileserver route is authenticated by the servlet chain
+before any application policy is consulted, so such a setting could only ever have described a
+permission the transport had already refused — a configuration that reads as if it grants access
+and does not.
+
+The three tiers do not inherit. An admin role cannot delete through the data plane, and a write role
+cannot reach the management plane — a role model where "can delete" implied "can force-delete" would
+make the audited plane reachable through the unaudited one.
+
+`unenforced` authorizes everything and exists so a developer can exercise upload and download before
+deciding on a role model. It is refused under a production profile.
+
+## Storage — `storage`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `root` | `/var/lib/backend/files` | absolute path; the only place a path exists |
+| `publish-mode` | `atomic-move-preferred` | `atomic-move-required`, `atomic-move-preferred`, `metadata-pointer` |
+| `buffer-size` | `128KB` | bounds every transfer allocation; memory never scales with file size |
+| `forbidden-root-ancestors` | `/app,/etc,/usr/share/nginx/html` | roots the storage root must not live under (webroot, config dirs) |
+
+`root` must be absolute. A relative root resolves against the process working directory, which is
+one path in a container and another in a test, so it is refused at binding time.
+
+Three former keys are gone, pinned as constants instead: staging and content share one FileStore,
+symbolic links are never followed, and the object and its directory are synced before READY. Each
+is an invariant the atomic publish and the namespace boundary are built on — a deployment that
+could switch one off would be running a different capability under the same name and the same
+tests.
+
+The storage provider has no selector either. There is exactly one implementation, and a `type` key
+with one legal value is a promise of pluggability that nothing keeps.
+
+## Upload, download, transfer
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `upload.max-file-size` | `100MB` | hard ceiling; also drives `spring.servlet.multipart.max-file-size` |
+| `upload.max-request-size` | `110MB` | request envelope; must be at least `max-file-size` |
+| `upload.initial-reservation` | `8MB` | quota reserved when the length is unknown |
+| `upload.ttl` | `1h` | how long a resumable upload stays claimable |
+| `upload.reservation-ttl` | `24h` | how long an unsettled quota reservation survives |
+| `upload.lease-duration` | `30s` | writer lease; renewed at one third of this |
+| `upload.max-parts` | `16` | multipart part ceiling |
+| `upload.require-content-length` | `false` | refuse chunked raw uploads |
+| `download.cache-control` | `private, no-store` | emitted on every content response |
+| `download.inline-allowed` | `false` | scriptable content is always an attachment regardless |
+| `download.max-ranges` | `1` | multi-range responses are opt-in |
+| `download.max-range-bytes` | `100MB` | total bytes one ranged response may cover |
+| `download.zero-copy-enabled` | `true` | hand large plaintext responses to the kernel |
+| `download.zero-copy-minimum-bytes` | `16MB` | below this the syscall setup costs more than it saves |
+| `transfer.core-size` / `max-size` / `queue-capacity` | `8` / `32` / `64` | blocking transfer pool bounds |
+| `transfer.await-seconds` | `300` | how long a transfer may occupy a pool thread |
+
+Zero copy changes no header and no status. When storage declines it — an unreadable region, an
+unsupported backend — the response is streamed instead and is byte-identical.
+
+## Verification — `verification`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `timeout` | `5s` | per-verifier ceiling |
+| `require-media-type-verdict` | `false` | refuse a file whose type could not be determined |
+| `inline-safe-profile` | `false` | accept scriptable content instead of quarantining it |
+
+Set `inline-safe-profile: true` only when downloads are never served inline from a trusted origin.
+
+## Quota and admission — `quota`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `instance-upload-permits` | `16` | concurrent uploads this node admits |
+| `scope-upload-permits` | `4` | concurrent uploads one namespace admits |
+| `direct-download-permits` | `64` | concurrent non-delegated downloads |
+| `soft-high-water` | `0.70` | storage fraction at which pressure is reported |
+| `hard-high-water` | `0.85` | storage fraction at which uploads are refused |
+
+When the storage fraction cannot be read, admission treats it as unknown and does not apply the
+high-water rule — a synthetic `0` would silently disable the guard, and a synthetic `1` would take
+the capability down over a failed syscall.
+
+## Background reclamation — `cleanup`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `enabled` | `false` | run the cleanup worker on this node |
+| `interval` | `60s` | fixed delay between batches, not fixed rate |
+| `max-items` | `100` | items one batch may claim |
+| `max-bytes` | `1GB` | bytes one batch may reclaim |
+| `retry-backoff` | `5m` | how long a failed item waits before it is due again |
+
+The worker deletes physical objects, so it is off until a deployment decides otherwise. A node
+without it still queues cleanup items; another node or an operator reclaims them. An item that fails
+eight times is abandoned rather than retried forever — it stays visible to an operator, parked
+rather than discarded.
+
+## Management plane — `admin`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `enabled` | `false` | expose `/internal/fileserver/**` |
+| `orphan-minimum-age` | `1h` | how long an unreferenced object must exist before a scan may name it |
+
+Publishing content and committing its record are two steps. Anything younger than
+`orphan-minimum-age` is assumed to be mid-commit rather than abandoned; shortening this makes
+concurrent uploads look like orphans.
+
+## Front-proxy delegation — `nginx`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `enabled` | `false` | emit `X-Accel-Redirect` instead of a body |
+| `internal-prefix` | `/__files/` | must be an `internal` location resolving to the content root |
+| `object-suffix` | `.bin` | layout suffix the proxy appends |
+| `minimum-size` | `16MB` | below this the application serves the transfer itself |
+
+Delegation is decided only after authorization and the READY gate, so an internal redirect can only
+ever name content the caller was already allowed to read.
+
+## Protocols — `tus`
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `enabled` | `false` | expose the tus 1.0 endpoints |
+
+The HTTPbis resumable-upload draft-12 surface is experimental and documented in
+[support-matrix.md](support-matrix.md).
@@ -0,0 +1,213 @@
+# Fileserver — deviations from the design
+
+The design specification and the implementation plan are frozen documents. Where implementation
+found them under-specified or self-contradicting, the resolution is recorded here rather than by
+editing the specification, and every entry names the test that pins the decision.
+
+## Resolved inconsistencies in the state machine and contracts
+
+### 1. `CREATED → FAILED` has no edge in the transition table
+
+A create that fails after the record exists must end in `FAILED`, but the table has no direct edge.
+The record therefore walks `CREATED → UPLOADING → FAILED`, which is also the honest reading: the
+upload had been admitted before it failed.
+
+Pinned by `UploadApplicationServiceTest` (application-core).
+
+### 2. `ContentKey`'s alphabet admits a leading separator
+
+The design's key pattern `[a-z0-9/_-]{16,200}` matches `/etc/passwd/...`. Rejecting an absolute path
+at the value type would change a design-fixed contract, so the stricter shape check lives in
+`PhysicalPathResolver`, per §12.2 rule 1 — the only place that turns an identifier into a path.
+
+Pinned by `PhysicalPathResolverTest`.
+
+### 3. `VERIFYING → DELETING` has no edge
+
+Deleting a file that is mid-verification has no legal transition. The lifecycle service refuses it
+with `409 FILE_NOT_READY` rather than inventing an edge, which matches the allowed-state list the
+JPA `markDeleting` statement already enforced.
+
+Pinned by `FileLifecycleServiceTest`.
+
+### 4. `If-Match` is specified as an ETag but the lifecycle was designed around the row version
+
+The HTTP contract sends an entity tag; the metadata store guards on a numeric version. The service
+takes `Optional<String> expectedEtag` and compares against the record's strong validator, so the
+precondition a client sends is the precondition that is checked.
+
+Pinned by `FileLifecycleServiceTest`.
+
+### 5. The filename policy left `:` intact
+
+`C:\Windows\system.ini` sanitized to `C:Windowssystem.ini` — a drive-qualified name surviving into
+display text and headers. `:` joined the structural strip set.
+
+Pinned by `FileserverHardeningContractTest` and `AmbiguousFilesystemOperationDetectorTest`.
+
+## Additions the design implies but does not specify
+
+### 6. `fs_recovery_item`
+
+§10.2 lists five core tables and none of them can hold the recovery queue, yet §29.3 requires one:
+reconciliation reports files whose bytes and metadata disagree, and holding that list in memory
+would lose exactly the cases a restart interrupted. Added in
+`V2__fileserver_recovery_and_staging_cleanup.sql` with one open item per file, so repeated sweeps
+update a worklist rather than accumulating a log.
+
+Pinned by `PostgreSqlFileserverReclamationIntegrationTest`.
+
+### 7. `fs_cleanup_item.upload_id`
+
+A staging object is addressed by upload, not by file. Without this column a queued staging cleanup
+could name only already-published content, so a cancelled or expired upload left bytes nothing could
+find. Added in the same migration, with a check constraint that an item names exactly one target.
+
+Pinned by `PostgreSqlFileserverReclamationIntegrationTest`.
+
+### 8. `ContentReferenceLedger` and `StagingUploadLocator`
+
+The orphan scan must ask whether a record still claims a physical object, and reconciliation must
+map a file back to the upload that last staged it. Neither question is answerable through the
+design's `FileMetadataStore` or `UploadSessionStore` as written. Rather than widen those
+design-fixed interfaces, both are narrow single-method ports.
+
+Pinned by `PostgreSqlFileserverReclamationIntegrationTest` and `LocalOrphanScanAdapterTest`.
+
+## Interpretations
+
+### 9. Quota settlement is FIFO within a scope
+
+Nothing links a reservation row to the upload that took it, and the design deliberately reclaims
+stragglers by TTL and the `STALE_QUOTA_RESERVATION` cleanup type rather than threading a reservation
+id through the upload session. `QuotaCommitGateway` therefore settles the oldest live reservation in
+the file's namespace.
+
+Which row closes does not change any quota decision: enforcement sums reserved and committed bytes
+per scope and never reads an individual row. Concurrent uploads of different sizes can leave the
+reserved total transiently high or low, and it converges as each settles. Durable usage with no live
+reservation behind it — an upload that outlived its TTL — is still recorded, because a ledger that
+silently under-counts is worse than one that is briefly imprecise.
+
+Pinned by `PostgreSqlFileserverReclamationIntegrationTest`.
+
+### 10. Zero copy is a channel transfer, not a file handoff
+
+Task 22 asks for zero copy on local files; §19 forbids a `Path` leaving the storage adapter, and §5
+of the plan forbids adding `Path` to the content store. WebFlux's zero-copy API takes a `Path`, so
+that route is closed.
+
+The servlet path takes the other one: `ZeroCopyDownloadGateway` receives a `WritableByteChannel` from
+the transport and the storage adapter performs `FileChannel.transferTo` into it. That is a genuine
+kernel-level transfer with no filesystem concept leaving storage. The reactive path continues to
+stream with bounded demand.
+
+Zero copy is an optimization with no observable difference: when storage declines, the response is
+streamed and is byte-identical.
+
+Pinned by `LocalStorageGatewayContractTest` and `ZeroCopyEligibilityTest`.
+
+### 11. The Fileserver JPA stores are gated on the capability switch
+
+The metadata store, session store, quota service, queues, ledger, and staging locator carry
+`@ConditionalOnProperty(app.fileserver-platform.enabled)` even though the rest of
+`adapter:outbound:persistence-jpa` is unconditional.
+
+Without the gate, every composition root that includes the persistence module built these beans —
+including `sample-portfolio`, which has no Fileserver — and each of them needs collaborators only
+the Fileserver configuration provides. That is the same rule the design states for the transport
+surface, applied to persistence: no surface appears merely because the dependency is present.
+
+`FileStateMachine` is bound alongside them, in `FileserverStorageConfiguration`. It had no
+production binding at all before, which made the metadata store unconstructible in any
+component-scanned context.
+
+Pinned by `SampleApplicationContextTest` (the capability off) and
+`FileserverRuntimeAssemblyTest` (the capability on).
+
+### 12. Transaction boundaries are owned by the application services, and are deliberately narrow
+
+The design does not say where a transaction begins. The repository does:
+`adapter:outbound:persistence-jpa` forbids a repository adapter from owning a `@Transactional`
+boundary, and `application-core` owns them through `TransactionPort`. The Fileserver follows that
+rule — every `Jpa*` store here declares no `@Transactional` of its own.
+
+What is specific to this capability is how narrow the boundaries are. A boundary covers a contiguous
+run of metadata writes and **stops before every storage call**, because a filesystem operation
+inside a database transaction would hold a connection for the length of a byte transfer. The upload
+path therefore has three boundaries, not one: acquire the lease, transfer the bytes, commit the
+offset.
+
+Where several stores must agree, they share one boundary:
+
+| Unit | Why it is one boundary |
+| --- | --- |
+| reserve quota + insert record + create session | a reservation that outlived a failed insert holds capacity for a file that never existed |
+| READY transition + quota commit | a finished file whose reservation was never converted holds capacity until the reservation expires |
+| `markDeleting` + enqueue cleanup | a file that stopped being reachable with nothing queued to reclaim it is never collected |
+| content delete settlement: reclaim + retire record + close queue item | half of it leaves the item to be retried against content that no longer exists |
+
+What this cannot make atomic is the storage/metadata seam itself — no database boundary could. That
+seam is exactly what the ambiguous-completion path and the reconciler exist for, and the one
+hand-written compensation that remains (staging creation failing after the records committed) is
+there for the same reason.
+
+Pinned by `FileserverRoundTripContractTest` against real PostgreSQL; the application tests use
+`DirectTransactions`, which runs a boundary inline and counts it.
+
+## Not implemented
+
+### `AsyncContentStore`, `CapacityAwareContentStore`, `CopyCapableContentStore`, `DelegatedDownloadStore`
+
+Four optional content-store SPIs are declared in `application-core` with no implementation. Each is
+an extension point for a backend this template does not ship:
+
+- `AsyncContentStore` — for a backend whose native client is non-blocking. The local platform is
+  blocking, and the reactive transport bridges to it on a dedicated I/O scheduler.
+- `CapacityAwareContentStore` — capacity is reported through `StorageHealthPort` and
+  `StorageUsageProbe`, which the local platform implements.
+- `CopyCapableContentStore` — server-side copy is delivered by `CopyContentGateway`; the local
+  platform has no cheaper primitive than a streamed copy.
+- `DelegatedDownloadStore` — delegation is delivered at the transport boundary by the nginx
+  `X-Accel-Redirect` strategy, which needs no store participation.
+
+`ContentStoreCapabilities` reports what the running store actually supports, so no unimplemented SPI
+is advertised as available.
+
+## Known deviation from the repository's application-layer contract
+
+### 5. Fileserver application services are not `CommandUseCase` / `QueryUseCase`
+
+`src/application-core/CLAUDE.md` requires every inbound port implementation to extend
+`CommandUseCase` or `QueryUseCase` and to carry `@UseCaseCapability`, which declares its transaction
+mode, idempotency and repository access. The Fileserver instead exposes multi-method services —
+`UploadApplicationService`, `DownloadApplicationService`, `FileLifecycleService`,
+`FileserverAdminService` and their `Default*` implementations.
+
+This is a real deviation, not an oversight, and it is unenforced: the ArchUnit rules
+`inbound_port_implementations_end_with_use_case` and
+`inbound_port_implementations_declare_capability` only match types that implement `UseCase`, so a
+service that never does is silently exempt. The capability contract that every other feature in
+this repository declares is therefore absent here.
+
+Two things follow from it. The transaction mode of each operation is expressed only by which
+`TransactionPort` method the body happens to call, rather than declared and checked. And the
+application layer holds transport policy it would not hold if each operation were a use case with
+its own command: HTTP status codes on `FileserverErrorCode`, `Range` and conditional-request
+parsing in `api.transfer`, and `Content-Disposition` construction.
+
+The status mapping in particular is a deliberate trade rather than an accident. It lives in
+`application-core` so the servlet transport, the reactive transport and the Nginx delegation path
+cannot answer the same failure with three different statuses. Moving it to the transport layer
+resolves the layering complaint and reintroduces exactly that drift, which is why this is an
+architecture decision rather than a cleanup.
+
+**Status: open, deliberately unresolved in this change set.** Closing it means roughly thirty
+command/query use cases, a decision about where the shared status vocabulary lives, and a change to
+the ArchUnit rules so a service that bypasses the contract fails the build instead of being exempt
+from it. That belongs in its own ADR with its own review, and doing it inside a correctness patch
+would mix a large mechanical refactor into changes that need to be readable.
+
+Nothing here is pinned by a test, because the deviation is the absence of a constraint. The next
+step is the ADR, not another test.
+
@@ -0,0 +1,105 @@
+# Fileserver HTTP contract
+
+Every public endpoint is listed here. `FileserverDocumentationCoverageTest` scans the controllers
+and fails if one is missing, so this file cannot silently fall behind the code.
+
+## Public endpoints
+
+| Method | Path | Success | Notes |
+|---|---|---|---|
+| POST | `/v1/files` | `201` READY, `202` VERIFYING | multipart single upload |
+| POST | `/v1/files:raw` | `201`, `202` | the whole request body is the file |
+| POST | `/v1/files:batch` | `200` | ordered per-part results; explicitly non-atomic |
+| GET | `/v1/files/{fileId}` | `200` | public metadata; never a content key or path |
+| GET | `/v1/files/{fileId}/content` | `200`, `206`, `304` | download |
+| HEAD | `/v1/files/{fileId}/content` | `200`, `304` | identical headers, no body |
+| DELETE | `/v1/files/{fileId}` | `202`, `204` | logical delete first |
+| POST | `/v1/files/{fileId}:copy` | `202` | create-only target |
+| POST | `/v1/files/{fileId}:move` | `200` | logical namespace change only |
+| OPTIONS | `/v1/uploads` | `204` | tus capability discovery |
+| POST | `/v1/uploads` | `201` | tus creation |
+| HEAD | `/v1/uploads/{uploadId}` | `204` | tus offset |
+| PATCH | `/v1/uploads/{uploadId}` | `204` | tus append |
+| DELETE | `/v1/uploads/{uploadId}` | `204` | tus termination |
+| POST | `/v1/experimental/draft12/uploads` | `201` | Experimental; off by default |
+| PATCH | `/v1/experimental/draft12/uploads/{uploadId}` | `204` | Experimental; off by default |
+
+## Management endpoints
+
+Reachable only where both `app.fileserver-platform.enabled=true` and
+`app.fileserver-platform.admin.enabled=true`, gated at the servlet chain on
+`app.fileserver-platform.security.admin-roles`, and intended for a management
+port rather than the public one.
+
+| Method | Path |
+|---|---|
+| GET | `/internal/fileserver/storage-health` |
+| GET | `/internal/fileserver/capabilities` |
+| GET | `/internal/fileserver/orphans` |
+| POST | `/internal/fileserver/orphans:reconcile` |
+| POST | `/internal/fileserver/files/{fileId}:reverify` |
+| POST | `/internal/fileserver/files/{fileId}:force-delete` |
+| GET | `/internal/fileserver/uploads/incomplete` |
+| POST | `/internal/fileserver/uploads:cleanup` |
+
+## Status codes
+
+| Status | Condition |
+|---:|---|
+| `200` | metadata, full GET, batch result, move |
+| `201` | file or upload created |
+| `202` | verification or physical cleanup deferred |
+| `204` | append, cancel, bodyless update |
+| `206` | satisfiable Range |
+| `304` | validator matched on GET or HEAD |
+| `400` | malformed header or header combination |
+| `401` | unauthenticated |
+| `403` / `404` | denied, or hidden under the existence-hiding profile |
+| `409` | state, offset, or lease conflict |
+| `410` | expired upload resource |
+| `411` | `require-content-length` profile with no length |
+| `412` | precondition failed |
+| `413` | size or quota policy violation |
+| `415` | upload media type not accepted |
+| `416` | unsatisfiable Range; carries the real length |
+| `422` | digest, signature, or scanner rejection |
+| `429` | transfer admission or rate limit |
+| `503` | storage or scanner unavailable |
+| `504` | downstream timeout |
+| `507` | out of storage capacity |
+
+## Failure body
+
+Every failure answers `application/problem+json` with a stable code and its URN:
+
+```json
+{
+  "type": "urn:fileserver:problem:upload-offset-mismatch",
+  "title": "Upload offset mismatch",
+  "status": 409,
+  "code": "UPLOAD_OFFSET_MISMATCH",
+  "retryable": true,
+  "ambiguous": false,
+  "reconciliationRequired": false,
+  "traceId": "..."
+}
+```
+
+The server-side exception message never appears. `ambiguous` is the field a client must read before
+retrying: an ambiguous failure may already have taken effect.
+
+## Header contract
+
+| Header | Contract |
+|---|---|
+| `Content-Type` | client value is a claim; the verified type is stored separately |
+| `Content-Disposition` | `attachment` by default; scriptable types are never inline |
+| `Accept-Ranges` | `bytes` |
+| `Range` | single range by default; multi-range only under an explicit budget |
+| `Content-Range` | actual range on `206`; the unsatisfied form on `416` |
+| `ETag` | strong validator derived from the SHA-256 |
+| `Last-Modified` | metadata publication instant, never a filesystem timestamp |
+| `Cache-Control` | `private, no-store` by default |
+| `X-Content-Type-Options` | always `nosniff` on a download |
+| `Retry-After` | on retryable `409`, `429`, `503`, and `504` |
+| `X-Accel-Redirect` | internal only; never forwarded to a client |
@@ -0,0 +1,128 @@
+# Fileserver runbooks
+
+Each runbook names the exact metric that fires it and the exact command that resolves it. A runbook
+whose trigger is "someone noticed" is not actionable, so every one below starts from a signal.
+
+## Storage full
+
+**Signal** — `fileserver.quota{result="rejected"}` rising, or `507` responses appearing.
+
+Storage capacity is exhausted or the high-water guard tripped. Uploads are rejected before any bytes
+are written, so nothing is corrupt; the system is refusing work it cannot complete.
+
+```bash
+curl -s $ADMIN/internal/fileserver/storage-health | jq '.usedFraction, .usableBytes'
+curl -s -X POST "$ADMIN/internal/fileserver/uploads:cleanup?maxItems=500&maxBytes=10737418240"
+curl -s "$ADMIN/internal/fileserver/orphans?limit=200" | jq '[.[].sizeBytes] | add'
+```
+
+Drain the cleanup backlog first — it reclaims space the system already knows is dead. Only then
+consider an orphan reconcile, and start with a dry run.
+
+## Orphan growth
+
+**Signal** — `fileserver.cleanup{result="skipped"}` climbing, or the orphan scan returning more
+objects each run.
+
+Physical objects exist with no metadata record pointing at them. This is not immediately dangerous —
+nothing serves them — but it consumes capacity indefinitely.
+
+```bash
+# Always look first. A reconcile without dryRun=false is a plan, not an action.
+curl -s -X POST "$ADMIN/internal/fileserver/orphans:reconcile" \
+  -H 'content-type: application/json' -d '{"limit":100}' | jq '.candidates'
+
+# Apply only the fingerprints you were just shown.
+curl -s -X POST "$ADMIN/internal/fileserver/orphans:reconcile" \
+  -H 'content-type: application/json' \
+  -d '{"dryRun":false,"limit":100,"maxBytes":1073741824,
+       "expectedFingerprints":["<from the dry run>"],"reasonCode":"ORPHAN_GROWTH_RUNBOOK"}'
+```
+
+Echoing the fingerprints is the safety property: an object that changed between the scan and the
+apply is skipped rather than deleted.
+
+## Verification backlog
+
+**Signal** — `fileserver.verification.queue{age_bucket="old"}` non-zero, or files sitting in
+VERIFYING.
+
+A verifier is slow or unavailable. Files stay non-public, which is the correct failure direction: a
+`RETRY` verdict never becomes an `ACCEPT`.
+
+```bash
+curl -s $ADMIN/internal/fileserver/capabilities | jq '.storageType'
+# Once the verifier is healthy, quarantined files can be re-examined individually.
+curl -s -X POST "$ADMIN/internal/fileserver/files/$FILE_ID:reverify"
+```
+
+Do not clear the backlog by disabling verification. A file that reached READY without an accepting
+verdict cannot be distinguished later from one that was verified.
+
+## NFS ambiguity
+
+**Signal** — problem documents carrying `"ambiguous": true`, or
+`fileserver.transfer.interruption{reason="stale_handle"}`.
+
+An operation's outcome could not be determined: the response was lost after the write or rename may
+have landed. These are never retried automatically.
+
+```bash
+# The recovery queue holds the files awaiting a decision.
+curl -s "$ADMIN/internal/fileserver/uploads/incomplete?limit=100" | jq
+```
+
+Reconciliation compares the physical size and digest against the record and only confirms READY when
+all four of key, size, digest, and version agree. Anything short of that is reported, never guessed.
+
+## PVC remount
+
+**Signal** — startup failure naming "atomic move", "same file store", or "not writable".
+
+The volume was remounted somewhere the probe can no longer prove a required capability. The
+application refuses traffic rather than serving from storage it cannot publish to atomically.
+
+```bash
+kubectl apply -f infra/fileserver/kubernetes/pvc-certification-job.yaml
+kubectl logs job/fileserver-pvc-certification
+```
+
+Compare the printed tuple with the certified one in `docs/fileserver/storage-certification.md`. A
+mismatch in CSI driver, StorageClass, access mode, or mount options is the cause; the certification
+does not carry across it.
+
+## Nginx delegation failure
+
+**Signal** — `fileserver.download.delegation{delegated="true"}` with client-visible `404`s.
+
+The internal location is misconfigured, so the proxy cannot resolve the redirect it was handed.
+
+```bash
+# The internal prefix must resolve to the content root and must be marked `internal`.
+grep -A5 '__files' infra/fileserver/nginx/nginx.conf
+curl -s $ADMIN/internal/fileserver/capabilities | jq '.capabilities.delegatedDownload'
+```
+
+Turning delegation off is a safe immediate mitigation: the application serves the transfer itself,
+slower but correct.
+
+```bash
+app.fileserver-platform.nginx.enabled=false
+```
+
+## Cleanup backlog
+
+**Signal** — `fileserver.cleanup{result="deferred"}` rising, or reclaimed bytes flat while deletes
+continue.
+
+Items are being deferred faster than they drain. The usual cause is an active writer lease still
+holding staging objects, which is correct behaviour, not a fault.
+
+```bash
+curl -s "$ADMIN/internal/fileserver/uploads/incomplete?limit=100" \
+  | jq '[.[] | select(.leaseUntil != null)] | length'
+curl -s -X POST "$ADMIN/internal/fileserver/uploads:cleanup?maxItems=500&maxBytes=10737418240"
+```
+
+If the deferrals are all `ACTIVE_WRITER_LEASE`, the backlog resolves itself as those uploads expire.
+Never delete staging content to clear a backlog: an upload that is mid-flight will corrupt.
@@ -0,0 +1,74 @@
+# Fileserver security model
+
+## The rule everything else follows
+
+Uploaded content is attacker-controlled. Every guard below exists because some part of the request
+— the filename, the declared media type, the range, the offset — is a value the caller chose.
+
+## Path safety
+
+A client value never becomes a path. The physical key is server-generated, and `ContentKey`'s
+character class excludes `.` entirely, so no traversal or extension-shaped segment survives
+validation. `DefaultPhysicalPathResolver` is the only place an identifier becomes a `Path`, and it
+normalizes and re-checks containment after construction rather than trusting the input.
+
+Symlink refusal happens at open time, not only at construction. A parent directory can be replaced
+between the two, so a check that ran only at path-building time would be a race, not a guard.
+
+## Filename handling
+
+`OriginalFilenamePolicy` strips path separators, NUL, quoting characters, and the colon — the last
+because on Windows it opens both a drive reference and an NTFS alternate data stream, so a name that
+keeps it is still path-shaped after the slashes are gone. Control characters and bidirectional
+overrides are removed, dot runs collapsed, reserved device names guarded, and the result is bounded
+in UTF-8 bytes.
+
+The sanitized name is display data. It is never used to build a key, and it reaches a header only
+through `ContentDispositionFactory`, which restricts the ASCII form and percent-encodes the UTF-8
+form.
+
+## Content type
+
+The client's `Content-Type` is stored as a claim. The verified type comes from the verification
+pipeline, and only the verified type is served. A claimed type that contradicts the content is
+quarantined rather than corrected.
+
+Scriptable types are never served inline, whatever the caller asked for: serving stored HTML or SVG
+inline from an upload origin is a stored cross-site scripting primitive. Every download also carries
+`X-Content-Type-Options: nosniff`.
+
+## Verification precedence
+
+`REJECT > QUARANTINE > RETRY > ACCEPT`. A verifier that times out or throws is `RETRY`, never a
+silent pass, and an empty verifier chain answers `RETRY` rather than accepting. A file becomes
+publicly readable only after an `ACCEPT`.
+
+## Range safety
+
+The range budget is enforced before content is opened, so a request naming many ranges is rejected
+without amplifying into storage work. An unsatisfiable range answers `416` with the real length and
+opens nothing.
+
+## Authorization
+
+Every public operation calls the injected `FileAccessPolicy` before any quota reservation or storage
+mutation, so a denial leaves no record, no reservation, and no staging object. Startup refuses to
+run a production profile with an allow-all policy.
+
+## Delegation
+
+`X-Accel-Redirect` is emitted only after authorization and the READY gate, and only for a full,
+unconditional response. The internal prefix must be an `internal` Nginx location; the front proxy
+also strips any client-supplied delegation header so a caller cannot name an internal object.
+
+## Telemetry
+
+No metric label, span attribute, or audit record carries a file id, upload id, filename, path, or
+user id. Where correlation is needed the value is a keyed HMAC fingerprint — keyed because the
+identifier space is enumerable and an unkeyed digest of it is reversible by brute force.
+
+## Ambiguous failures
+
+A failure whose operation may already have taken effect is reported as ambiguous and is never
+retryable. On a network filesystem a lost response is indistinguishable from a rejection at the
+socket level, so anything not provably safe is treated as ambiguous and sent to reconciliation.
@@ -0,0 +1,66 @@
+# Storage certification
+
+## Why a certification is per-volume
+
+Atomic rename, same-file-store guarantees, and symlink refusal are properties of a specific
+filesystem behind a specific mount — not of "Kubernetes" or "a PVC". Change the CSI driver, the
+StorageClass, the access mode, the backend, or the mount options and any of them can differ. A
+certification that does not name all five is not transferable.
+
+## What is certified
+
+| Property | Why it matters |
+|---|---|
+| Same file store for staging and content | A rename across stores is a copy, so publication stops being atomic. |
+| Atomic rename | The publish path's default strategy. |
+| Atomic create (`O_EXCL`) | Makes a publish create-only rather than a silent overwrite. |
+| Symlink refusal | Stops a replaced parent from redirecting a write outside the root. |
+| Ranged read | The download contract depends on it. |
+
+## Running the certification
+
+```bash
+kubectl apply -f infra/fileserver/kubernetes/pvc-certification-job.yaml
+kubectl logs job/fileserver-pvc-certification
+```
+
+The job writes a machine-readable result to the claim itself, carrying the full tuple:
+
+```json
+{
+  "kubernetesVersion": "...",
+  "csiDriver": "...",
+  "storageClass": "...",
+  "accessMode": "ReadWriteOnce",
+  "backend": "ext2/ext3",
+  "mountOptions": "rw,relatime",
+  "atomicMove": true,
+  "sameFileStore": true,
+  "atomicCreate": true
+}
+```
+
+The job fails closed: a volume whose staging and content areas are on different stores is not
+certified, because its publish would silently degrade to a copy.
+
+## Network filesystems
+
+```bash
+docker compose -f infra/fileserver/nfs/compose.yml up -d
+FILESERVER_NFS_TESTS=true ./gradlew :adapter:outbound:fileserver:test
+```
+
+The mount is `hard`, deliberately. A `soft` mount converts a slow server into a short write, which
+is exactly the corruption this design refuses to accept.
+
+## Startup enforcement
+
+`FileserverStartupValidator` re-runs the probe at boot and refuses to accept traffic when a required
+capability is missing — `ATOMIC_MOVE_REQUIRED` on a filesystem that cannot prove an atomic move
+fails closed rather than degrading silently.
+
+## Adding a new store
+
+Extend `ContentStoreContract` and pass it. A prose claim of compatibility is not accepted; the
+contract is executable precisely so a future object-storage adapter has to demonstrate the same
+offset, digest, and create-only behaviour the local store does.
@@ -0,0 +1,83 @@
+# Fileserver support matrix
+
+A support level here is a claim about evidence, not about intent. Every row names the CI job that
+produces that evidence; `FileserverDocumentationCoverageTest` fails the build if a row names a job
+that does not exist, so a level can never outlive the test that justified it.
+
+## Levels
+
+| Level | What it means |
+|---|---|
+| Stable | Certified on every pull request. Contract changes are breaking changes. |
+| Beta | Certified nightly. The contract may still change with a deprecation notice. |
+| Limited | Certified on the release gate only, under stated constraints. |
+| Compatibility | Accepted but not optimized; known caveats are listed inline. |
+| Experimental | Off by default, unratified upstream, may change without notice. |
+
+## Runtime profiles
+
+| Profile | Level | CI job |
+|---|---|---|
+| Local filesystem (ext4) content store | Stable | `fileserver-local-ext4-contract` |
+| Spring MVC transport (raw, multipart, batch, download) | Stable | `fileserver-http-contract` |
+| Spring WebFlux transport | Experimental | `fileserver-http-contract` |
+| Path, filename, range, and problem-detail hardening | Stable | `fileserver-security-suite` |
+| Bounded-memory transfer | Stable | `fileserver-bounded-memory` |
+| Application and architecture invariants | Stable | `fileserver-unit-and-architecture` |
+| Runtime assembly (the capability starts with the flag on) | Stable | `fileserver-unit-and-architecture` |
+| tus 1.0 resumable uploads | Stable | `fileserver-http-contract` |
+| Crash-recovery matrix | Beta | `fileserver-process-kill-matrix` |
+| NFSv4 ambiguity handling | Beta | `fileserver-nfs-ambiguity` |
+| Large-file and slow-client performance | Beta | `fileserver-large-file-performance` |
+| Multi-instance writer lease | Beta | `fileserver-multi-instance-lease` |
+| Kubernetes ReadWriteOnce PVC | Limited | `fileserver-pvc-certification` (manifest checks in CI; cluster run is operator-driven) |
+| Nginx `X-Accel-Redirect` delegation | Limited | `fileserver-http-contract` |
+| Telemetry sensitive-data suppression | Stable | `fileserver-sensitive-telemetry-scan` |
+| Documentation and support-claim coverage | Stable | `fileserver-documentation-gate` |
+| Full release verification | Stable | `fileserver-full-verification` |
+| HTTP resumable uploads draft-12 | Experimental | `fileserver-http-contract` |
+
+### Why WebFlux is Experimental, not Stable
+
+The reactive router, handlers and readers are now wired: `FileserverReactiveConfiguration`
+contributes the scheduler, the handlers and a `RouterFunction` bean under
+`@ConditionalOnWebApplication(type = REACTIVE)` plus the platform master switch. Previously nothing
+built them at all, so "Stable" described the source tree rather than a running server.
+
+It stays `Experimental` because the shipped composition cannot select it. `adapter:inbound:web`
+also puts `DispatcherServlet` on the classpath — deliberately, so adding `spring-webflux` does not
+drag a second embedded server onto the runtime — and Boot's application-type deduction therefore
+resolves SERVLET. A fork that removes the servlet stack and adds a reactive server gets working
+routes without editing any Fileserver code; the shipped template does not exercise that path.
+
+Raising it to Stable requires a contract job that drives the routes over a running reactive server
+rather than through direct construction.
+
+## Explicitly not claimed
+
+These have no job, and therefore no claim:
+
+- An automated Kubernetes cluster result. `fileserver-pvc-certification` validates the manifest on
+  every release and applies it only when a release cluster is configured; without one it warns and
+  records that nothing was certified. The cluster tuple is produced by an operator and read from
+  [storage-certification.md](storage-certification.md).
+
+- Kubernetes ReadWriteMany PVC. Concurrent writers across nodes are not certified.
+- Windows NTFS as a production storage root. The filename policy strips the characters NTFS
+  reserves, but no job certifies the publish path there.
+- Object storage as a content store. The contract exists (`ContentStoreContract`) but no adapter
+  implements it yet.
+- Server-side malware scanning. The verification pipeline has the port and the verdict precedence;
+  no scanner is shipped.
+
+## Where the rest is written down
+
+- [configuration.md](configuration.md) — every `app.fileserver-platform.*` key, its default, and the
+  conditions that fail startup rather than degrade.
+- [design-deviations.md](design-deviations.md) — where the implementation departs from the frozen
+  design, why, and the test that pins each decision.
+- [http-contract.md](http-contract.md) — the wire contract.
+- [security.md](security.md) — the threat model and what enforces each control.
+- [operations.md](operations.md) — runbooks, each starting from a metric.
+- [storage-certification.md](storage-certification.md) — how a volume is certified.
+- [upgrade-guide.md](upgrade-guide.md) — what changes between versions.
@@ -0,0 +1,82 @@
+# Fileserver upgrade guide
+
+## Enabling the capability
+
+The Fileserver ships off. Nothing is registered — no endpoint, no thread pool, no metric — until it
+is enabled explicitly.
+
+```yaml
+ca-skeleton:
+  fileserver:
+    enabled: true
+    instance-id: ${HOSTNAME}
+    default-namespace: default
+    observability:
+      fingerprint-key: ${FILESERVER_FINGERPRINT_KEY}
+```
+
+`instance-id` must be unique per instance: it is the writer-lease owner, and two nodes sharing one
+would both believe they hold the same lease.
+
+`fingerprint-key` is required and has no default. Startup fails without it rather than falling back
+to an unkeyed digest, which would be reversible for an enumerable identifier space.
+
+## Optional surfaces
+
+Each is a separate switch, and each defaults to off:
+
+```yaml
+ca-skeleton:
+  fileserver:
+    admin:
+      enabled: false        # management plane; intended for a management port
+    tus:
+      enabled: false        # tus 1.0 Stable
+    httpbis-draft12:
+      enabled: false        # Experimental; unratified, may change without notice
+    nginx:
+      enabled: false        # front-proxy delegation; needs a validated internal location
+```
+
+## Database schema
+
+The metadata schema is installed as a capability migration and starts inactive:
+
+```
+V1__create_fileserver_metadata.sql  →  capability_schema_registry: jpa-fileserver-metadata-v1
+```
+
+Activate it deliberately. Enabling the capability without an activated schema fails at startup
+rather than at the first upload.
+
+## Choosing a publish mode
+
+| Mode | When |
+|---|---|
+| `atomic-move-preferred` | Default. Uses an atomic rename when the probe proves one, else a metadata pointer. |
+| `atomic-move-required` | Fail closed. Refuses to start on storage that cannot prove an atomic move. |
+| `metadata-pointer` | For storage without atomic rename; publication is the metadata commit. |
+
+Pick `atomic-move-required` when the storage is certified and you want a misconfiguration to surface
+at boot rather than at publish time.
+
+## Behaviour that will surprise you
+
+- **A delete answers `202`, not `204`, when content still exists.** The file is already unreadable;
+  the physical reclaim is deferred. Treating `202` as a failure will produce spurious retries.
+- **A batch upload answers `200` even when parts failed.** The batch is explicitly non-atomic, and a
+  single status could not report a partial outcome honestly. Read `results[].problem`.
+- **An ambiguous failure must not be retried.** Check `"ambiguous": true` in the problem document.
+- **`If-Match` takes the strong ETag, not a version number.** A client can only assert about the
+  representation it was actually served.
+- **Inline rendering is refused for scriptable types** even when the caller asks for it.
+
+## Verifying an upgrade
+
+```bash
+cd src
+./gradlew verifyCleanArchitectureDependencies --console=plain
+./gradlew :application-core:check :adapter:inbound:web:check \
+          :adapter:outbound:fileserver:check --console=plain
+./gradlew :app-bootstrap:test --tests '*Fileserver*' --console=plain
+```
@@ -0,0 +1,6 @@
+44ba9931722364a53fcb3b5f31a1d539eabcaf42db775f5a33fb558f558c7504  README.md
+d064f0ac6c3be0e5c76ef22454db2a97e1d78ed287bd22f4c125f19aba3ad8e3  VALIDATION.md
+1ef15812f33dc998a6332b87523ed5942ba46d79d984a0ca776b05bb9247a06a  docs/superpowers/plans/2026-08-12-graphql-advanced-capabilities-expansion-plan.md
+5ae70b53e22cdb852b2bb0df171dec868bfe99b15bb8e71fb2b0b3431cd7e2cd  docs/superpowers/plans/2026-08-12-graphql-api-execution-platform-implementation-plan.md
+8d0203203f6bfe4b2e18625eff23bb308ba6454703a4ca4cd3236dab31ecafc3  docs/superpowers/specs/2026-08-12-graphql-api-execution-platform-design.md
+8048fe6a536de67d2cf5b0df05d35128f2c68ba8f0dd615831b40430fc76277b  validate_graphql_docs.py
@@ -0,0 +1,43 @@
+# GraphQL Superpowers 설계 패키지
+
+이 패키지는 `GraphQL API 실행 플랫폼 심층 리서치`를 구현 기준선으로 변환한 설계서와 실행 계획서다.
+
+## 문서
+
+- `docs/superpowers/specs/2026-08-12-graphql-api-execution-platform-design.md`
+  - Stable·Advanced 전체 아키텍처, 공개 계약, 경계, 실패 의미론, 테스트와 지원 등급
+  - 입력 심층 리서치 원문을 추적 부록으로 포함
+- `docs/superpowers/plans/2026-08-12-graphql-api-execution-platform-implementation-plan.md`
+  - Stable 구현 Task 1–48
+- `docs/superpowers/plans/2026-08-12-graphql-advanced-capabilities-expansion-plan.md`
+  - Stable Release Gate 이후 실행하는 Advanced·Experimental Task 1–19
+- `VALIDATION.md`
+  - 정적 검증 결과와 검증 범위
+- `validate_graphql_docs.py`
+  - 패키지 내부 문서 재검증 스크립트
+- `MANIFEST.sha256`
+  - 패키지 파일 무결성 목록
+
+## 구현 순서
+
+```text
+Stable Task 1–48
+→ Stable Release Gate
+→ Advanced Task 1–19
+→ Capability별 Promotion Gate
+```
+
+## 명시적 전제
+
+```text
+Java 21
+Gradle Kotlin DSL
+Spring Boot 4.1 BOM
+Spring for GraphQL 2.0
+Boot-managed GraphQL Java v25 계열
+Stable module root: modules/graphql
+Advanced module root: modules/graphql-advanced
+Root package: io.backend.skeleton.graphql
+```
+
+실제 저장소에 적용할 때 기존 package·version catalog·module naming에 맞춰 경로만 조정하고, 문서의 공개 계약·불변 조건·테스트 의미는 유지한다.
@@ -0,0 +1,98 @@
+# GraphQL Superpowers 문서 정적 검증 결과
+
+- **검증 시각 기준:** 2026-08-12
+- **검증 대상:** 설계서 1개, Stable 구현 계획서 1개, Advanced·Experimental 확장 계획서 1개
+- **검증 명령:** `python3 validate_graphql_docs.py`
+- **결과:** **PASS**
+- **실행 검사:** 1,475
+- **통과:** 1,475
+- **실패:** 0
+
+## 문서 규모
+
+| 문서 | 행 수 | 크기 |
+|---|---:|---:|
+| GraphQL API 실행 플랫폼 설계서 | 2,553 | 93,359 bytes |
+| Stable 구현 계획서 | 4,560 | 209,041 bytes |
+| Advanced 확장 계획서 | 1,976 | 105,717 bytes |
+
+## 계획 구조
+
+| 항목 | Stable | Advanced |
+|---|---:|---:|
+| Task 수 | 48 | 19 |
+| Create 경로 수 | 227 | 113 |
+| Task 번호 연속성 | PASS | PASS |
+| 모든 Task의 `Files`·`Interfaces` | PASS | PASS |
+| 모든 Task의 Implementation Requirements | PASS | PASS |
+| 모든 Task의 Step 1–5 | PASS | PASS |
+| 실패·통과 예상 결과 | PASS | PASS |
+| Task별 Git commit 명령 | PASS | PASS |
+| Create 경로 중복 | 없음 | 없음 |
+| Stable·Advanced 경로 충돌 | 없음 | 없음 |
+
+## 핵심 계약 검증
+
+```text
+SDL-first external contract
+Single Executable Schema Stable default
+HTTP POST Stable profile
+application/graphql-response+json preferred
+Validation 이후 Field Error·Partial Data는 HTTP 200
+Draft 294는 Stable에서 제외
+JPA Entity·MongoDB Document 직접 노출 금지
+GraphQL Multipart Upload 미지원·Fileserver 사용
+request-wide database transaction 금지
+DataLoader request scope
+Finite Fetch Profile
+HMAC-signed cursor
+Mutation idempotency·expected version 분리
+Parser·shape·complexity·runtime response budget
+Actor·Field·Object·Tenant authorization
+Low-cardinality observability
+Stable/Advanced dependency isolation
+Persisted Operation·WebSocket·SSE·Federation 분리
+RSocket·HTTP GET·Incremental Delivery Experimental
+```
+
+위 계약은 설계서와 계획서의 필수 문자열·모듈 경로·Task별 파일·테스트를 대조해 검증했습니다.
+
+## 입력 리서치 추적성
+
+- 첨부된 `GraphQL API 실행 플랫폼 심층 리서치` 원문 전체가 설계서의 `부록 B`에 포함되어 있습니다.
+- 설계 본문은 원문의 용어와 결론을 유지하면서 구현 판단을 Stable·Advanced·Experimental로 고정합니다.
+- 설계서와 입력 원문의 exact text 포함 검사를 별도로 통과했습니다.
+
+## 패키지 검증 항목
+
+```text
+문서 파일 존재
+Markdown code fence 균형
+Task 1–48 / 1–19 연속성
+Task별 테스트·명령·commit
+정확한 Create 경로
+Placeholder 금지
+Stable module에 WebSocket·Federation·Persisted Operation 경로 부재
+Advanced module에 feature flag와 capability 경로 존재
+금지 API pattern 부재
+문서 SHA-256 계산
+```
+
+## 검증 범위의 한계
+
+현재 PASS는 **문서의 정적 구조, 요구사항 추적성, 내부 계약과 실행 계획의 완결성**을 의미합니다. 실제 Backend Skeleton 저장소가 입력으로 제공되지 않았으므로 다음은 실행하지 않았습니다.
+
+```text
+Gradle configuration·compile
+Spring Boot ApplicationContext 기동
+SchemaMappingInspector 실제 결과
+GraphQlTester HTTP·WebFlux contract
+JPA·MongoDB statement/query-count integration
+query bomb·complexity load test
+Virtual Thread·event-loop blocking test
+WebSocket·SSE soak test
+Federation composition·router integration
+actual Git commit
+```
+
+실제 구현에서는 Stable Task 1–48을 먼저 수행해 Stable Release Gate를 통과한 뒤 Advanced Task 1–19를 시작해야 합니다.
@@ -0,0 +1,249 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+from pathlib import Path
+import re
+import sys
+import hashlib
+
+ROOT = Path(__file__).resolve().parent
+DESIGN = ROOT / "docs/superpowers/specs/2026-08-12-graphql-api-execution-platform-design.md"
+STABLE = ROOT / "docs/superpowers/plans/2026-08-12-graphql-api-execution-platform-implementation-plan.md"
+ADVANCED = ROOT / "docs/superpowers/plans/2026-08-12-graphql-advanced-capabilities-expansion-plan.md"
+
+checks: list[tuple[str, bool, str]] = []
+
+def check(name: str, condition: bool, detail: str = "") -> None:
+    checks.append((name, bool(condition), detail))
+
+def read(path: Path) -> str:
+    check(f"file exists: {path.name}", path.exists(), str(path))
+    return path.read_text(encoding="utf-8") if path.exists() else ""
+
+design = read(DESIGN)
+stable = read(STABLE)
+advanced = read(ADVANCED)
+
+# Basic document integrity
+check("design line floor", len(design.splitlines()) >= 2000, str(len(design.splitlines())))
+check("stable plan line floor", len(stable.splitlines()) >= 4000, str(len(stable.splitlines())))
+check("advanced plan line floor", len(advanced.splitlines()) >= 1500, str(len(advanced.splitlines())))
+for label, text in [("design", design), ("stable", stable), ("advanced", advanced)]:
+    check(f"{label} code fences balanced", text.count("```") % 2 == 0, str(text.count("```")))
+    for marker in ["TODO", "TBD", "FIXME", "implement later", "fill in details"]:
+        check(f"{label} no placeholder {marker}", marker.lower() not in text.lower())
+
+# Design required sections and source traceability
+required_design_terms = [
+    "# GraphQL API 실행 플랫폼 설계서",
+    "GraphQL Platform owns",
+    "Domain/Application owns",
+    "G1 Standard GraphQL API",
+    "G2 Advanced Execution",
+    "G3 GraphQL Extension",
+    "G4 Admin Plane",
+    "SDL",
+    "September 2025",
+    "application/graphql-response+json",
+    "HTTP `200`",
+    "GraphQlRequestContext",
+    "DataLoader",
+    "GraphQlFetchProfile",
+    "HMAC",
+    "Idempotency",
+    "Partial Data",
+    "Persisted Operation",
+    "Subscription",
+    "Federation",
+    "GraphQL Multipart Upload",
+    "Fileserver",
+    "부록 B. 입력 심층 리서치 원문",
+    "# GraphQL API 실행 플랫폼 심층 리서치",
+]
+for term in required_design_terms:
+    check(f"design contains {term}", term in design)
+
+# Critical design invariants
+critical_pairs = [
+    ("field error uses HTTP 200", "field error" in design.lower() and "HTTP `200`" in design),
+    ("no draft 294 stable", "294" in design and "Stable" in design),
+    ("dataloader request scope", "request" in design.lower() and "DataLoader" in design),
+    ("cursor HMAC", "Cursor" in design and "HMAC" in design),
+    ("no multipart upload", "Multipart Upload" in design and "Fileserver" in design),
+    ("single schema default", "Single Executable Schema" in design),
+    ("request-wide transaction prohibited", "request-wide" in design.lower() and "transaction" in design.lower()),
+    ("entity/document boundary", "JPA Entity" in design and "MongoDB Document" in design),
+]
+for name, condition in critical_pairs:
+    check(name, condition)
+
+# Plan headers and global constraints
+stable_header_terms = [
+    "# GraphQL API 실행 플랫폼 Implementation Plan",
+    "REQUIRED SUB-SKILL",
+    "**Goal:**",
+    "**Architecture:**",
+    "**Tech Stack:**",
+    "## Global Constraints",
+    "Stable Task",
+]
+advanced_header_terms = [
+    "# GraphQL Advanced Capability Expansion Implementation Plan",
+    "REQUIRED SUB-SKILL",
+    "backend.graphql.advanced.*",
+    "Stable 구현 계획 Task `1–48`",
+]
+for term in stable_header_terms:
+    check(f"stable header contains {term}", term in stable)
+for term in advanced_header_terms:
+    check(f"advanced header contains {term}", term in advanced)
+
+# Task sequence and per-task structure
+def task_sections(text: str) -> list[tuple[int, str]]:
+    matches = list(re.finditer(r"^### Task (\d+): .+$", text, re.MULTILINE))
+    result = []
+    for i, match in enumerate(matches):
+        start = match.start()
+        end = matches[i+1].start() if i+1 < len(matches) else len(text)
+        result.append((int(match.group(1)), text[start:end]))
+    return result
+
+stable_tasks = task_sections(stable)
+advanced_tasks = task_sections(advanced)
+check("stable task count", len(stable_tasks) == 48, str(len(stable_tasks)))
+check("advanced task count", len(advanced_tasks) == 19, str(len(advanced_tasks)))
+check("stable task sequence", [n for n, _ in stable_tasks] == list(range(1, 49)))
+check("advanced task sequence", [n for n, _ in advanced_tasks] == list(range(1, 20)))
+
+def validate_tasks(label: str, tasks: list[tuple[int, str]]) -> None:
+    required = [
+        "**Files:**",
+        "**Interfaces:**",
+        "**Implementation requirements:**",
+        "**Step 1: Write the failing test**",
+        "**Step 2: Run the focused test and verify the failure**",
+        "**Step 3: Implement the smallest complete production contract**",
+        "**Step 4: Run the focused test and the owning suite**",
+        "**Step 5: Commit the independently reviewable change**",
+        "Expected: FAIL",
+        "Expected: PASS",
+        "git commit -m",
+    ]
+    for number, section in tasks:
+        for token in required:
+            check(f"{label} task {number} contains {token}", token in section)
+        check(f"{label} task {number} has test path", "- Test: `" in section)
+        check(f"{label} task {number} has production file", "- Create: `" in section)
+        check(f"{label} task {number} fences balanced", section.count("```") % 2 == 0)
+        check(f"{label} task {number} has gradle test", "./gradlew" in section and ":test" in section)
+
+validate_tasks("stable", stable_tasks)
+validate_tasks("advanced", advanced_tasks)
+
+# Create paths
+def create_paths(text: str) -> list[str]:
+    return re.findall(r"^- Create: `([^`]+)`$", text, re.MULTILINE)
+
+stable_paths = create_paths(stable)
+advanced_paths = create_paths(advanced)
+check("stable create paths exist", len(stable_paths) >= 150, str(len(stable_paths)))
+check("advanced create paths exist", len(advanced_paths) >= 80, str(len(advanced_paths)))
+check("stable create paths unique", len(stable_paths) == len(set(stable_paths)))
+check("advanced create paths unique", len(advanced_paths) == len(set(advanced_paths)))
+check("stable and advanced paths disjoint", set(stable_paths).isdisjoint(advanced_paths))
+for index, path in enumerate(stable_paths, 1):
+    check(f"stable create path {index} exact", "*" not in path and "..." not in path and (path.startswith("modules/graphql/") or path.startswith("build-logic/")))
+for index, path in enumerate(advanced_paths, 1):
+    check(f"advanced create path {index} exact", "*" not in path and "..." not in path and path.startswith("modules/graphql-advanced/"))
+
+# Stable/Advanced separation
+for forbidden in [
+    "modules/graphql/graphql-websocket/",
+    "modules/graphql/graphql-federation/",
+    "modules/graphql/graphql-persisted-operation/",
+    "modules/graphql/graphql-rsocket/",
+]:
+    check(f"stable excludes {forbidden}", forbidden not in stable)
+
+for required in [
+    "modules/graphql-advanced/graphql-persisted-operation/",
+    "modules/graphql-advanced/graphql-websocket/",
+    "modules/graphql-advanced/graphql-subscription/",
+    "modules/graphql-advanced/graphql-federation/",
+    "modules/graphql-advanced/graphql-rsocket/",
+]:
+    check(f"advanced includes {required}", required in advanced)
+
+# Stable coverage
+stable_required_terms = [
+    "GraphQlRequestContext",
+    "GraphQlClientPolicy",
+    "GraphQlSchemaContract",
+    "SchemaMappingInspector",
+    "@oneOf",
+    "GraphQlHttpProfile",
+    "application/graphql-response+json",
+    "GraphQlExecutionProfile",
+    "GraphQlWireError",
+    "GraphQlTenantIsolationPolicy",
+    "GraphQlParserLimits",
+    "GraphQlComplexityCalculator",
+    "GraphQlRuntimeBudget",
+    "GraphQlPreparsedCacheKey",
+    "GraphQlBatchPolicy",
+    "GraphQlFetchProfile",
+    "HmacGraphQlCursorCodec",
+    "GraphQlConnection",
+    "GraphQlMutationIdempotencyContext",
+    "GraphQlMetricCardinalityPolicy",
+    "GraphQlPlatformStartupValidator",
+    "GraphQlReleaseGate",
+]
+for term in stable_required_terms:
+    check(f"stable coverage {term}", term in stable)
+
+advanced_required_terms = [
+    "GraphQlPersistedOperation",
+    "GraphQlWebSocketProtocol",
+    "GraphQlSubscriptionBufferPolicy",
+    "GraphQlSubscriptionOrderingProfile",
+    "GraphQlSseConnectionPolicy",
+    "GraphQlReplayPosition",
+    "GraphQlDataLoaderDependencyGraph",
+    "GraphQlFederationEntityKey",
+    "GraphQlFederationCompositionGate",
+    "GraphQlGeneratedSourceBoundary",
+    "GraphQlRepositoryAllowlist",
+    "GraphQlRSocketRoutePolicy",
+    "GraphQlHttpGetOperationPolicy",
+    "GraphQlIncrementalCompatibilityGate",
+    "GraphQlAdvancedReleaseGate",
+]
+for term in advanced_required_terms:
+    check(f"advanced coverage {term}", term in advanced)
+
+# Prohibited API patterns
+prohibited_patterns = [
+    (r"interface\s+GenericGraphQlRepository", "no generic graphql repository"),
+    (r"public\s+.*\bEntityManager\b", "no public entity manager"),
+    (r"public\s+.*\bMongoTemplate\b", "no public mongo template"),
+    (r"scalar\s+Upload\b", "no upload scalar declaration"),
+    (r"@Transactional\s+.*GraphQL request", "no request-wide transaction implementation"),
+]
+for pattern, name in prohibited_patterns:
+    check(name, re.search(pattern, stable, re.IGNORECASE | re.MULTILINE) is None)
+
+# File hashes can be printed for package evidence
+for path in [DESIGN, STABLE, ADVANCED]:
+    if path.exists():
+        digest = hashlib.sha256(path.read_bytes()).hexdigest()
+        check(f"sha256 computed: {path.name}", len(digest) == 64, digest)
+
+failed = [(n, d) for n, ok, d in checks if not ok]
+print(f"CHECKS={len(checks)}")
+print(f"PASSED={len(checks)-len(failed)}")
+print(f"FAILED={len(failed)}")
+for name, detail in failed:
+    print(f"FAIL: {name}" + (f" :: {detail}" if detail else ""))
+
+sys.exit(1 if failed else 0)
@@ -0,0 +1,228 @@
+# HTTP Client Platform — Configuration Reference
+
+Every outbound call resolves exactly one **Named Client Profile**. The whole capability lives under
+the `app.httpclient` prefix: profiles under `app.httpclient.clients[N]`, Dynamic Target policies
+under `app.httpclient.dynamic-targets[N]`.
+
+Design §30.1 forbids a production profile from inheriting large framework defaults. Anything a
+production deployment must decide has either no default or an unusable one, and
+`HttpClientStartupValidator` fails the context rather than guessing.
+
+## The master switch
+
+| Property | Type | Default | Environment |
+|---|---|---|---|
+| `app.httpclient.enabled` | boolean | `false` | `APP_HTTPCLIENT_ENABLED` |
+
+Off is the shipped state and it is a structural one. `HttpClientPlatformAutoConfiguration` lives in
+a package the composition root's component scan excludes, so while the switch is absent or false the
+class is never processed and neither is anything it imports: no property is bound, and no transport
+provider, connection pool, TLS context, credential, thread, gateway or actuator endpoint exists. A
+malformed HTTP client setting cannot fail the startup of a deployment that never wanted outbound
+HTTP.
+
+Anything that is not exactly `true` — `yes`, `1`, blank — leaves the platform off. Turning it on
+with no client declared is a startup failure carrying `HTTPCLIENT_ACTIVE_WITHOUT_CLIENTS`: a
+platform with nothing to call still holds transport providers and gateways no caller can reach.
+
+## Declaring clients from the environment
+
+Clients are an indexed list carrying their own `name`, not a map keyed by name. A map key becomes a
+segment of the environment variable and the relaxed binder normalises it, so `payment-api` and
+`payment_api` would arrive as one entry with nothing said about the one that was lost. Both a
+duplicate name and a name that collides once normalised fail startup.
+
+```dotenv
+APP_HTTPCLIENT_ENABLED=true
+
+APP_HTTPCLIENT_CLIENTS_0_NAME=payment
+APP_HTTPCLIENT_CLIENTS_0_BASE_URL=https://payment.example
+APP_HTTPCLIENT_CLIENTS_0_ALLOWED_HOSTS_0=payment.example
+APP_HTTPCLIENT_CLIENTS_0_ALLOWED_PORTS_0=443
+APP_HTTPCLIENT_CLIENTS_0_REQUEST_MAX_BODY_BYTES=1048576
+APP_HTTPCLIENT_CLIENTS_0_TLS_PROFILE_ID=payment
+
+APP_HTTPCLIENT_DYNAMIC_TARGETS_0_NAME=webhook
+APP_HTTPCLIENT_DYNAMIC_TARGETS_0_ALLOWED_SCHEMES_0=https
+```
+
+`docs/httpclient/env-fields.yaml` is the registry of accepted variable names. It is
+derived from the settings record and held to it in both directions, and the platform refuses to
+start on an `APP_HTTPCLIENT_` variable that is not in it — so
+`APP_HTTPCLIENT_CLIENTS_0_TIMEUOT_TOTAL_CALL` fails startup instead of silently leaving the client
+on its default budget. Unknown keys supplied through a configuration file rather than the
+environment are refused by strict binding for the same reason.
+
+Only `APP_HTTPCLIENT_ENABLED` appears in `src/.env` and `docs/registries/env-keys.yaml`. It is the
+one key with a deployment-independent value; templating an indexed client in `application.yml` would
+materialise a nameless client in every deployment, which the aggregate validation refuses.
+
+## `app.httpclient.clients[N]`
+
+| Property | Type | Default | Notes |
+|---|---|---|---|
+| `name` | string | — | Required, unique, and distinct from every other name once normalised for the environment |
+| `mode` | `TRUSTED` \| `DYNAMIC` | `TRUSTED` | A dynamic profile may not carry a default credential |
+| `base-url` | URI | — | Required for a trusted profile; no userinfo, no query |
+| `allowed-hosts` | list | empty | Required in production |
+| `allowed-ports` | list | empty | Compared against the effective port |
+| `api` | `REST_CLIENT` \| `WEB_CLIENT` | `REST_CLIENT` | Decides blocking or reactive runtime |
+| `transport` | `APACHE` \| `JDK` \| `REACTOR_NETTY` \| `JETTY` \| `SIMPLE` | `APACHE` | `SIMPLE` is rejected in production |
+| `protocols` | list | `HTTP_1_1` | The default transport is Apache, whose classic client is HTTP/1.1 only; a profile that wants HTTP/2 declares it together with a transport that can deliver it. `HTTP_3` requires the experimental acknowledgement |
+| `experimental-acknowledgement` | string | — | Must equal `I_ACCEPT_HTTP3_EXPERIMENTAL_SEMANTICS` |
+
+### `pool`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `max-total-connections` | `50` | Socket ceiling for the runtime |
+| `max-connections-per-route` | `25` | Per-upstream ceiling |
+| `max-pending-acquires` | `100` | Waiting-request memory ceiling |
+| `pending-acquire-timeout` | `200ms` | Pool or stream wait ceiling |
+| `max-idle-time` | `30s` | Idle eviction |
+| `max-life-time` | `5m` | Picks up DNS, load-balancer, and certificate changes |
+| `validate-after-inactivity` | `5s` | Stale and half-open detection |
+| `eviction-interval` | `15s` | Background cleanup |
+| `shutdown-timeout` | `5s` | Drain deadline before forced close |
+| `requires-route-pool` | `false` | Set when route-scoped limits are mandatory; the JDK transport then refuses the profile |
+| `requires-bounded-pending-queue` | `false` | Same, for a bounded pending queue |
+
+### `timeout`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `dns` | `300ms` | Hostname resolution |
+| `connect` | `500ms` | Socket connect |
+| `tls-handshake` | `1s` | TLS and ALPN |
+| `proxy-connect` | `500ms` | Proxy socket or CONNECT |
+| `request-write-idle` | `1s` | No progress writing the request |
+| `response-header` | `2s` | Until final response headers |
+| `read-idle` | `3s` | Between response chunks |
+| `total-call` | `4s` | The whole logical call, including retry backoff |
+| `streaming-idle` | `30s` | Silence on a long-lived stream |
+
+`total-call` must not be shorter than `connect` or `response-header`; the validator emits
+`INVALID_TIMEOUT_BUDGET` otherwise.
+
+### `redirect`, `request`, `response`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `redirect.enabled` | `false` | Engine redirect handling is always off; the platform follows hops itself |
+| `redirect.max-hops` | `0` | Enabling redirects with zero hops is a configuration error |
+| `redirect.allow-cross-origin` | `false` | When enabled, credentials are stripped on the hop |
+| `request.max-body-bytes` | `0` | Required in production |
+| `request.compression` | `false` | |
+| `response.max-wire-bytes` | `5242880` | Bytes on the wire |
+| `response.max-decoded-bytes` | `10485760` | Bytes after decoding; hard maximum is 64 MiB |
+| `response.allowed-content-types` | JSON + problem+json | Empty means "any" |
+
+### `authentication`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `type` | `NONE` | One of the design §20.1 methods |
+| `registration-id` | — | Required for OAuth2 |
+| `scopes` | empty | Part of the token cache key |
+| `audience` | — | Part of the token cache key |
+| `header-name` | — | Required for `API_KEY_HEADER`; must be on the allowlist |
+| `secret-reference` | — | Resolved by the deployment's secret loader, never a literal |
+
+### `retry`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `policy` | `none` | Named policy for reporting |
+| `max-attempts` | `1` | Attempts, not retries |
+| `base-backoff` | `50ms` | |
+| `max-backoff` | `200ms` | |
+| `jitter` | `FULL` | `NONE` \| `FULL` \| `DECORRELATED` |
+| `retry-after` | `HONOR` | `HONOR` \| `IGNORE` \| `CAP` |
+| `budget` | — | Shared token bucket name |
+
+### `tls`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `profile-id` | — | Required in production; the only TLS identifier the actuator exposes |
+| `protocols` | `TLSv1.3, TLSv1.2` | Anything else is rejected |
+| `hostname-verification` | `true` | Setting it false fails startup |
+| `trust-all` | `false` | Exists only so the unsafe intent is rejectable; nothing acts on `true` |
+| `allow-plain-http` | `false` | Plaintext fallback fails startup in production |
+| `trust-material-reference` | — | Custom CA, resolved by the secret loader |
+| `key-material-reference` | — | Client certificate for mTLS |
+
+### `proxy` and `observability`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `proxy.enabled` | `false` | |
+| `proxy.host` / `proxy.port` / `proxy.type` | — / `0` / `HTTP` | |
+| `proxy.credential-provider` | — | Proxy authentication is separate from target authentication |
+| `proxy.connect-timeout` | `500ms` | Recorded as its own metric |
+| `proxy.import-ambient-no-proxy` | `false` | Ambient `NO_PROXY` never widens a validated profile |
+| `observability.operation-name-required` | `true` | |
+| `observability.full-url-recording` | `false` | |
+| `observability.body-logging` | `false` | |
+
+## `app.httpclient.dynamic-targets[N]`
+
+| Property | Default | Meaning |
+|---|---|---|
+| `name` | — | Required, unique, and subject to the same normalisation rule as a client name |
+| `allowed-schemes` | `https` | |
+| `allowed-ports` | `443` | |
+| `allowed-host-suffixes` | empty | |
+| `allowed-hosts` | empty | Empty means "any host that survives address validation" |
+| `max-redirect-hops` | `0` | Each hop repeats the full validation flow |
+| `trace-propagation` | `false` | Off by default for dynamic targets |
+| `blocked-cidrs` | empty | Organisation-defined internal ranges |
+
+## Startup violation codes
+
+`TRUSTED_BASE_URL_REQUIRED`, `BASE_URL_USERINFO_FORBIDDEN`, `BASE_URL_QUERY_FORBIDDEN`,
+`PLAINTEXT_PRODUCTION_TARGET`, `ALLOWED_HOST_MISMATCH`, `ALLOWED_PORT_MISMATCH`,
+`REDIRECT_POLICY_INVALID`, `REDIRECT_CROSS_ORIGIN_CREDENTIAL_POLICY_REQUIRED`,
+`INVALID_TIMEOUT_BUDGET`, `RESPONSE_HARD_MAXIMUM_EXCEEDED`, `PRODUCTION_SIMPLE_FACTORY_FORBIDDEN`,
+`JDK_FINE_GRAINED_POOL_UNSUPPORTED`, `HTTP3_STABLE_FORBIDDEN`,
+`DYNAMIC_TARGET_TRANSPORT_UNSUPPORTED`, `DYNAMIC_DEFAULT_CREDENTIAL_FORBIDDEN`,
+`OAUTH2_REGISTRATION_REQUIRED`, `API_KEY_HEADER_NAME_REQUIRED`, `TRUST_ALL_FORBIDDEN`,
+`HOSTNAME_VERIFICATION_REQUIRED`, `PLAINTEXT_FALLBACK_FORBIDDEN`, `TLS_PROTOCOL_FORBIDDEN`,
+`RETRY_BACKOFF_REQUIRED`, `MISSING_PRODUCTION_SETTING`, `DUPLICATE_CLIENT_NAME`,
+`HTTPCLIENT_ACTIVE_WITHOUT_CLIENTS`, `DYNAMIC_BASE_URL_REQUIRED`,
+`DYNAMIC_TARGET_PROXY_UNSUPPORTED`, `REACTIVE_AUTHENTICATION_UNSUPPORTED`,
+`HTTP2_REQUIRED_TRANSPORT_UNSUPPORTED`, `POOL_ROUTE_EXCEEDS_TOTAL`,
+`TLS_PROTOCOL_SET_REQUIRED`, `REACTIVE_REDIRECT_UNSUPPORTED`,
+`RETRY_POLICY_CONTRADICTS_ATTEMPTS`, `FULL_URL_RECORDING_FORBIDDEN`, `BODY_LOGGING_FORBIDDEN`,
+`DNS_TIMEOUT_UNSUPPORTED`, `PROXY_CREDENTIAL_UNSUPPORTED`, `PROXY_AMBIENT_NO_PROXY_UNSUPPORTED`.
+
+The last three name settings the platform binds but cannot yet honour. Neither the Apache classic
+client nor the JDK client exposes a DNS-resolution timeout, and no proxy-credential path exists, so
+a non-default value is refused rather than accepted and ignored. Leaving the defaults alone is
+unaffected — only a deliberate, unmet request fails.
+
+Three of these are about a guarantee that used to be silently unmet rather than refused:
+
+- `HTTP2_REQUIRED_TRANSPORT_UNSUPPORTED` — declaring `protocols: [HTTP_2]` alone states that HTTP/2
+  is required. Only `REACTOR_NETTY` can be configured to offer H2 and nothing else; the JDK client
+  treats it as a preference and negotiates HTTP/1.1, and Apache's classic client is HTTP/1.1 only.
+- `POOL_ROUTE_EXCEEDS_TOTAL` — a per-route ceiling above the total is incoherent, and on Reactor,
+  where the per-route knob is the only one that exists, it silently becomes the effective limit.
+- `TLS_PROTOCOL_SET_REQUIRED` — an empty `tls.protocols` used to pass and then let the JVM choose,
+  so emptying the list to "tighten" a profile loosened it.
+- `REACTIVE_REDIRECT_UNSUPPORTED` — engine redirect following is disabled on every transport and
+  only the blocking stack has a coordinator that follows hops with per-hop re-validation. A
+  `WEB_CLIENT` profile with `redirect.enabled=true` did not follow redirects; the caller received the
+  3xx as an ordinary response. Refused until the reactive coordinator exists.
+- `RETRY_POLICY_CONTRADICTS_ATTEMPTS` — `retry.policy` was read by nothing on the execution path, so
+  the actuator could report `none` for a profile retrying three times. The two settings must now
+  agree: `policy: none` requires `max-attempts: 1`, and any other policy requires more than one.
+- `FULL_URL_RECORDING_FORBIDDEN` / `BODY_LOGGING_FORBIDDEN` — both settings were bindable and inert.
+  Recording an expanded URL puts path identifiers and query strings into unbounded metric tags;
+  recording bodies puts someone else's data into logs. Representable so the intent is rejectable,
+  refused under a production profile.
+
+`DYNAMIC_TARGET_PROXY_UNSUPPORTED` is worth spelling out: a forward proxy resolves the hostname on
+its own side, so the addresses this platform validated and pinned are not the addresses the
+connection reaches. The SSRF defence would be present, correct, and bypassed — so the combination is
+refused rather than served with a guarantee it cannot keep.
@@ -0,0 +1,179 @@
+# HTTP Client platform — Java field path to environment variable template.
+#
+# The SSOT is HttpClientPlatformSettings. HttpClientEnvironmentKeys derives this list from the
+# record tree at runtime, HttpClientPlatformEnvManifestTest fails when the two disagree in either
+# direction, and the platform refuses to start on an APP_HTTPCLIENT_ variable that is not here. So a
+# field added with no entry, an entry whose field was renamed, and a misspelled variable in a
+# deployment are all failures rather than silence.
+#
+# `N` and `M` are list indices, not literals: `N` for the outermost list, `M` for a list inside it.
+# `app.httpclient.clients[N].base-url` is set as APP_HTTPCLIENT_CLIENTS_0_BASE_URL for the first
+# client, and `clients[N].allowed-hosts[M]` as APP_HTTPCLIENT_CLIENTS_0_ALLOWED_HOSTS_0.
+#
+# Only APP_HTTPCLIENT_ENABLED is registered in docs/registries/env-keys.yaml and shipped in
+# src/.env: it is the only key with a deployment-independent value, and it is the only one the
+# three-way :app-bootstrap:verifyEnvKeys gate can express. Everything below is per deployment and is set directly
+# in the environment — templating an indexed client in application.yml would materialise a nameless
+# client in every deployment, which the settings' aggregate validation refuses.
+#
+# This file lives beside the HTTP Client documentation rather than in docs/registries, which is a
+# fail-closed catalog of exactly eight contract registries with a fixed row schema
+# (owner_branch/compatibility_impact/required_test per row). A field-to-variable mapping does not
+# have that shape, and admitting it would have meant loosening a gate rather than satisfying one.
+#
+# Secrets are referenced, never carried: authentication.secret-reference, tls.*-material-reference
+# and proxy.credential-provider name material that a secret backend resolves. Putting the material
+# itself in one of these variables defeats the indirection they exist for.
+fields:
+  - field: enabled
+    env: APP_HTTPCLIENT_ENABLED
+  - field: clients[N].name
+    env: APP_HTTPCLIENT_CLIENTS_N_NAME
+  - field: clients[N].mode
+    env: APP_HTTPCLIENT_CLIENTS_N_MODE
+  - field: clients[N].base-url
+    env: APP_HTTPCLIENT_CLIENTS_N_BASE_URL
+  - field: clients[N].allowed-hosts[M]
+    env: APP_HTTPCLIENT_CLIENTS_N_ALLOWED_HOSTS_M
+  - field: clients[N].allowed-ports[M]
+    env: APP_HTTPCLIENT_CLIENTS_N_ALLOWED_PORTS_M
+  - field: clients[N].api
+    env: APP_HTTPCLIENT_CLIENTS_N_API
+  - field: clients[N].transport
+    env: APP_HTTPCLIENT_CLIENTS_N_TRANSPORT
+  - field: clients[N].protocols[M]
+    env: APP_HTTPCLIENT_CLIENTS_N_PROTOCOLS_M
+  - field: clients[N].pool.max-total-connections
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_MAX_TOTAL_CONNECTIONS
+  - field: clients[N].pool.max-connections-per-route
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_MAX_CONNECTIONS_PER_ROUTE
+  - field: clients[N].pool.max-pending-acquires
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_MAX_PENDING_ACQUIRES
+  - field: clients[N].pool.pending-acquire-timeout
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_PENDING_ACQUIRE_TIMEOUT
+  - field: clients[N].pool.max-idle-time
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_MAX_IDLE_TIME
+  - field: clients[N].pool.max-life-time
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_MAX_LIFE_TIME
+  - field: clients[N].pool.validate-after-inactivity
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_VALIDATE_AFTER_INACTIVITY
+  - field: clients[N].pool.eviction-interval
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_EVICTION_INTERVAL
+  - field: clients[N].pool.shutdown-timeout
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_SHUTDOWN_TIMEOUT
+  - field: clients[N].pool.requires-route-pool
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_REQUIRES_ROUTE_POOL
+  - field: clients[N].pool.requires-bounded-pending-queue
+    env: APP_HTTPCLIENT_CLIENTS_N_POOL_REQUIRES_BOUNDED_PENDING_QUEUE
+  - field: clients[N].timeout.dns
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_DNS
+  - field: clients[N].timeout.connect
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_CONNECT
+  - field: clients[N].timeout.tls-handshake
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_TLS_HANDSHAKE
+  - field: clients[N].timeout.proxy-connect
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_PROXY_CONNECT
+  - field: clients[N].timeout.request-write-idle
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_REQUEST_WRITE_IDLE
+  - field: clients[N].timeout.response-header
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_RESPONSE_HEADER
+  - field: clients[N].timeout.read-idle
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_READ_IDLE
+  - field: clients[N].timeout.total-call
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_TOTAL_CALL
+  - field: clients[N].timeout.streaming-idle
+    env: APP_HTTPCLIENT_CLIENTS_N_TIMEOUT_STREAMING_IDLE
+  - field: clients[N].redirect.enabled
+    env: APP_HTTPCLIENT_CLIENTS_N_REDIRECT_ENABLED
+  - field: clients[N].redirect.max-hops
+    env: APP_HTTPCLIENT_CLIENTS_N_REDIRECT_MAX_HOPS
+  - field: clients[N].redirect.allow-cross-origin
+    env: APP_HTTPCLIENT_CLIENTS_N_REDIRECT_ALLOW_CROSS_ORIGIN
+  - field: clients[N].request.max-body-bytes
+    env: APP_HTTPCLIENT_CLIENTS_N_REQUEST_MAX_BODY_BYTES
+  - field: clients[N].request.compression
+    env: APP_HTTPCLIENT_CLIENTS_N_REQUEST_COMPRESSION
+  - field: clients[N].response.max-wire-bytes
+    env: APP_HTTPCLIENT_CLIENTS_N_RESPONSE_MAX_WIRE_BYTES
+  - field: clients[N].response.max-decoded-bytes
+    env: APP_HTTPCLIENT_CLIENTS_N_RESPONSE_MAX_DECODED_BYTES
+  - field: clients[N].response.allowed-content-types[M]
+    env: APP_HTTPCLIENT_CLIENTS_N_RESPONSE_ALLOWED_CONTENT_TYPES_M
+  - field: clients[N].authentication.type
+    env: APP_HTTPCLIENT_CLIENTS_N_AUTHENTICATION_TYPE
+  - field: clients[N].authentication.registration-id
+    env: APP_HTTPCLIENT_CLIENTS_N_AUTHENTICATION_REGISTRATION_ID
+  - field: clients[N].authentication.scopes[M]
+    env: APP_HTTPCLIENT_CLIENTS_N_AUTHENTICATION_SCOPES_M
+  - field: clients[N].authentication.audience
+    env: APP_HTTPCLIENT_CLIENTS_N_AUTHENTICATION_AUDIENCE
+  - field: clients[N].authentication.header-name
+    env: APP_HTTPCLIENT_CLIENTS_N_AUTHENTICATION_HEADER_NAME
+  - field: clients[N].authentication.secret-reference
+    env: APP_HTTPCLIENT_CLIENTS_N_AUTHENTICATION_SECRET_REFERENCE
+  - field: clients[N].retry.policy
+    env: APP_HTTPCLIENT_CLIENTS_N_RETRY_POLICY
+  - field: clients[N].retry.max-attempts
+    env: APP_HTTPCLIENT_CLIENTS_N_RETRY_MAX_ATTEMPTS
+  - field: clients[N].retry.base-backoff
+    env: APP_HTTPCLIENT_CLIENTS_N_RETRY_BASE_BACKOFF
+  - field: clients[N].retry.max-backoff
+    env: APP_HTTPCLIENT_CLIENTS_N_RETRY_MAX_BACKOFF
+  - field: clients[N].retry.jitter
+    env: APP_HTTPCLIENT_CLIENTS_N_RETRY_JITTER
+  - field: clients[N].retry.retry-after
+    env: APP_HTTPCLIENT_CLIENTS_N_RETRY_RETRY_AFTER
+  - field: clients[N].retry.budget
+    env: APP_HTTPCLIENT_CLIENTS_N_RETRY_BUDGET
+  - field: clients[N].observability.operation-name-required
+    env: APP_HTTPCLIENT_CLIENTS_N_OBSERVABILITY_OPERATION_NAME_REQUIRED
+  - field: clients[N].observability.full-url-recording
+    env: APP_HTTPCLIENT_CLIENTS_N_OBSERVABILITY_FULL_URL_RECORDING
+  - field: clients[N].observability.body-logging
+    env: APP_HTTPCLIENT_CLIENTS_N_OBSERVABILITY_BODY_LOGGING
+  - field: clients[N].tls.profile-id
+    env: APP_HTTPCLIENT_CLIENTS_N_TLS_PROFILE_ID
+  - field: clients[N].tls.protocols[M]
+    env: APP_HTTPCLIENT_CLIENTS_N_TLS_PROTOCOLS_M
+  - field: clients[N].tls.hostname-verification
+    env: APP_HTTPCLIENT_CLIENTS_N_TLS_HOSTNAME_VERIFICATION
+  - field: clients[N].tls.trust-all
+    env: APP_HTTPCLIENT_CLIENTS_N_TLS_TRUST_ALL
+  - field: clients[N].tls.allow-plain-http
+    env: APP_HTTPCLIENT_CLIENTS_N_TLS_ALLOW_PLAIN_HTTP
+  - field: clients[N].tls.trust-material-reference
+    env: APP_HTTPCLIENT_CLIENTS_N_TLS_TRUST_MATERIAL_REFERENCE
+  - field: clients[N].tls.key-material-reference
+    env: APP_HTTPCLIENT_CLIENTS_N_TLS_KEY_MATERIAL_REFERENCE
+  - field: clients[N].proxy.enabled
+    env: APP_HTTPCLIENT_CLIENTS_N_PROXY_ENABLED
+  - field: clients[N].proxy.host
+    env: APP_HTTPCLIENT_CLIENTS_N_PROXY_HOST
+  - field: clients[N].proxy.port
+    env: APP_HTTPCLIENT_CLIENTS_N_PROXY_PORT
+  - field: clients[N].proxy.type
+    env: APP_HTTPCLIENT_CLIENTS_N_PROXY_TYPE
+  - field: clients[N].proxy.credential-provider
+    env: APP_HTTPCLIENT_CLIENTS_N_PROXY_CREDENTIAL_PROVIDER
+  - field: clients[N].proxy.connect-timeout
+    env: APP_HTTPCLIENT_CLIENTS_N_PROXY_CONNECT_TIMEOUT
+  - field: clients[N].proxy.import-ambient-no-proxy
+    env: APP_HTTPCLIENT_CLIENTS_N_PROXY_IMPORT_AMBIENT_NO_PROXY
+  - field: clients[N].experimental-acknowledgement
+    env: APP_HTTPCLIENT_CLIENTS_N_EXPERIMENTAL_ACKNOWLEDGEMENT
+  - field: dynamic-targets[N].name
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_NAME
+  - field: dynamic-targets[N].allowed-schemes[M]
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_ALLOWED_SCHEMES_M
+  - field: dynamic-targets[N].allowed-ports[M]
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_ALLOWED_PORTS_M
+  - field: dynamic-targets[N].allowed-host-suffixes[M]
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_ALLOWED_HOST_SUFFIXES_M
+  - field: dynamic-targets[N].allowed-hosts[M]
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_ALLOWED_HOSTS_M
+  - field: dynamic-targets[N].max-redirect-hops
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_MAX_REDIRECT_HOPS
+  - field: dynamic-targets[N].trace-propagation
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_TRACE_PROPAGATION
+  - field: dynamic-targets[N].blocked-cidrs[M]
+    env: APP_HTTPCLIENT_DYNAMIC_TARGETS_N_BLOCKED_CIDRS_M
@@ -0,0 +1,64 @@
+# Migrating from `RestTemplate`
+
+`RestTemplate` is permitted only inside `…httpclient.migration`; `RestTemplateBoundaryTest` enforces
+that. New retry, Dynamic Target, and HTTP/3 capabilities are deliberately unreachable from the
+migration path — a caller that wants them moves to a Named Client Profile.
+
+## 1. Audit before changing anything
+
+```java
+RestTemplateInventory inventory = new RestTemplateInventoryScanner().scan(existingTemplate);
+```
+
+The inventory reports the request factory, message converters, interceptors, error handler, and URI
+template handler, plus findings:
+
+| Code | Severity | Meaning |
+|---|---|---|
+| `SIMPLE_REQUEST_FACTORY` | blocking | no connection pool; unsupported in production |
+| `NO_MESSAGE_CONVERTERS` | blocking | the template cannot encode or decode a body |
+| `NO_INTERCEPTORS` | warning | confirm where correlation and timeouts are applied |
+| `TIMEOUTS_NOT_INTROSPECTABLE` | informational | declare timeouts explicitly on the target profile |
+
+## 2. Bridge without changing behaviour
+
+```java
+RestClient client = new RestTemplateToRestClientAdapter().adaptChecked(existingTemplate);
+```
+
+`adaptChecked` refuses to migrate a template with a blocking finding. The bridge carries the
+existing converters, interceptors, error handler, and URI handler across, so this step changes the
+API and nothing else.
+
+## 3. Move to a Named Client Profile
+
+Turn the platform on with `APP_HTTPCLIENT_ENABLED=true` — it ships off, and while it is off none of
+the settings below are bound — then declare the upstream as `app.httpclient.clients[N]` with its
+`name` and an explicit base URL, transport, timeouts, pool, body limits, authentication, retry
+policy, redirect policy, and TLS profile. Startup validation will tell you exactly which of those is
+missing. See `docs/httpclient/configuration-reference.md` for the environment form.
+
+## 4. Move to a typed client
+
+```java
+@HttpClientProfile("payment")
+@HttpExchange("/payments")
+public interface PaymentClient {
+
+  @PostExchange
+  @HttpOperationPolicy(
+      name = "create-payment",
+      idempotency = OperationIdempotency.IDEMPOTENCY_KEY_REQUIRED,
+      retryPolicy = "payment-write")
+  PaymentResponse create(
+      @RequestHeader("Idempotency-Key") String idempotencyKey, @RequestBody PaymentRequest request);
+}
+```
+
+The interface fails startup validation unless it declares a profile, gives every method a stable
+operation name and an explicit idempotency, supplies a key parameter when the operation requires
+one, keeps a single execution model, and does not enable retry on a non-idempotent write.
+
+## 5. Retire the template
+
+Once no production package references `RestTemplate`, `RestTemplateBoundaryTest` keeps it that way.
@@ -0,0 +1,121 @@
+# HTTP Client Platform — Operations Runbook
+
+## Metrics
+
+| Metric | Meaning |
+|---|---|
+| `http.client.requests` | Physical attempt timer (Spring standard name, kept deliberately) |
+| `http.client.logical.calls` | User-visible logical call timer |
+| `http.client.attempts` | Attempt counter |
+| `http.client.retry.count` | Retries by reason |
+| `http.client.retry.exhausted` | Retry budget exhausted |
+| `http.client.ambiguous` | Ambiguous outcomes |
+| `http.client.timeout` | Timeouts by stage |
+| `http.client.request.bytes` | Request wire bytes |
+| `http.client.response.bytes` | Response bytes |
+| `http.client.active` | In-flight attempts |
+| `http.client.pool.connections` | Leased and available connections |
+| `http.client.pool.pending` | Pool waiters |
+| `http.client.pool.acquire.duration` | Pool wait time |
+| `http.client.dns.duration` | DNS time |
+| `http.client.connect.duration` | Connect time |
+| `http.client.tls.duration` | TLS time |
+| `http.client.circuit.state` | Circuit state |
+| `http.client.bulkhead.rejected` | Bulkhead rejections |
+| `http.client.rate_limit.rejected` | Local rate-limit rejections |
+| `http.client.oauth.refresh` | Token refresh outcomes |
+| `http.client.ssrf.rejected` | Dynamic target rejections |
+
+`http.client.requests` counts attempts and `http.client.logical.calls` counts user calls. When they
+diverge, retries are absorbing failures — which is the first thing to look at during an incident.
+
+## Reading an incident
+
+| Symptom | Likely cause | Where to look |
+|---|---|---|
+| logical calls fine, attempts spiking | upstream degraded, retries absorbing it | `http.client.retry.count` by reason |
+| `http.client.ambiguous` non-zero | non-idempotent writes reaching `SENT_NO_RESPONSE` | reconcile with the upstream; consider an idempotency key |
+| pool pending climbing | pool too small or upstream slow | `http.client.pool.acquire.duration`, `pool.connections` |
+| circuit open | sustained upstream failure | `http.client.circuit.state`; local rejections do not open it |
+| `http.client.ssrf.rejected` non-zero | a caller is submitting internal URLs | Dynamic Target policy and audit trail |
+
+## Actuator
+
+`GET /actuator/httpclients` reports profile name, runtime generation, state, transport, API,
+protocols, active leases, pool ceiling, credential type, TLS profile id, redirect flag, retry policy,
+and capability warnings. Base URL, credentials, trust store paths, and resolved IPs are deliberately
+absent: an actuator endpoint is reachable by more people than a secret store is.
+
+## Rotation
+
+Certificates and secrets rotate by building a new runtime generation and swapping the registry
+pointer, never by mutating a live client. A connection pool holds sockets established under the
+previous identity, so replacing material without replacing the pool leaves live connections
+authenticated by a certificate that is meant to be gone.
+
+```text
+build new generation → validate → atomic swap → new calls use it
+old generation → DRAINING → in-flight calls finish → no new retries → forced close at the drain deadline
+```
+
+## Shutdown
+
+```text
+RUNNING → DRAINING
+new logical calls refused or routed to the new generation
+in-flight attempts complete
+new retries refused
+shutdown timeout
+remaining calls cancelled
+pool closed
+```
+
+## Retry ownership
+
+Exactly one of the application client, an external SDK, or the service mesh may own retries.
+Two owners multiply traffic during an incident. Record the owner per upstream and check it whenever
+a mesh retry policy changes.
+
+## Error model
+
+Every outbound failure is one of these stable types. The type is derived from the classified failure
+category, not from whatever the engine happened to throw, so it means the same thing on Apache, JDK,
+and Reactor Netty. Each carries `HttpFailureMetadata`: client, operation, method, URI **template**,
+evidence, replayability, stage, retryability, attempt, elapsed, remaining deadline, status, trace id
+— and nothing else.
+
+| Exception | Raised when | Retryable |
+|---|---|---|
+| `HttpConfigurationException` | profile, operation, or capability configuration is invalid | never |
+| `HttpTargetRejectedException` | target URI, host, port, header, or address policy refused the request | never |
+| `HttpDnsException` | hostname resolution failed or timed out | yes, inside budget |
+| `HttpPoolAcquireTimeoutException` | no connection or stream within the pending-acquire budget | yes, inside budget |
+| `HttpConnectException` | socket connect failed | yes, inside budget |
+| `HttpProxyException` | proxy connect, CONNECT tunnel, or proxy auth failed | yes, inside budget |
+| `HttpTlsException` | TLS handshake failed | only a transient handshake timeout |
+| `HttpRequestWriteException` | request headers or body could not be fully written | only when safely idempotent |
+| `HttpResponseTimeoutException` | final headers or a body chunk did not arrive in time | only when safely idempotent |
+| `HttpResponseTruncatedException` | the response ended before the body was complete | only when safely idempotent and undelivered |
+| `HttpRemoteErrorException` | non-success status without a problem document | per the status rules |
+| `HttpProblemDetailException` | non-success status with a bounded RFC 9457 document | per the status rules |
+| `HttpRedirectRejectedException` | a hop violated hop count, origin, method, or replay policy | never |
+| `HttpAuthenticationException` | credential materialization or refresh failed | never |
+| `HttpSerializationException` | request encoding or response decoding failed | never |
+| `HttpResponseTooLargeException` | wire or decoded bytes exceeded the profile limit | never |
+| `HttpDeadlineExceededException` | the effective deadline was reached | never |
+| `HttpCircuitOpenException` | the upstream circuit is open | never |
+| `HttpBulkheadRejectedException` | no attempt or logical admission permit was available | never |
+| `HttpRateLimitRejectedException` | the local attempt rate limit or retry budget rejected the attempt | never |
+| `HttpAmbiguousExecutionException` | a non-idempotent request was sent and the outcome is unknown | never — reconcile instead |
+
+## Traces
+
+```text
+http.client.operation        logical internal span
+└─ http.client.request       attempt 1 CLIENT span
+└─ http.client.request       attempt 2 CLIENT span
+```
+
+W3C Trace Context is propagated with a Baggage allowlist. Dynamic Targets do not propagate trace
+context by default. Retry reason and evidence are recorded as span events; credentials and remote
+error bodies are never recorded as attributes.
@@ -0,0 +1,55 @@
+# HTTP Client Platform — Performance Baseline
+
+The certification lane asserts **resource bounds**, not throughput targets. Its purpose is to prove
+that a failing upstream, a large body, or a rotation cannot consume unbounded memory, connections,
+threads, or upstream traffic. Nothing here becomes a runtime adaptive default: every bound comes
+from an explicit profile setting.
+
+## How to run
+
+```bash
+# structural bounds only (default; still executes every test)
+./gradlew :adapter:outbound:httpclient:httpClientPerformanceTest --console=plain
+
+# full certification, including machine-dependent bounds
+./gradlew :adapter:outbound:httpclient:httpClientPerformanceTest \
+  -Pperformance.assertions.enabled=true --console=plain
+
+# JMH benchmarks
+./gradlew :adapter:outbound:httpclient:jmh --console=plain
+```
+
+Machine-dependent assertions are reported as explicitly skipped when the flag is absent — the lane
+never silently degrades into a pass.
+
+## Certified bounds
+
+| Test | Bound | Kind |
+|---|---|---|
+| `RetryStormBudgetTest` | 10 000 logical calls against a failing upstream produce at most 11 000 physical attempts at a 10 % budget | structural |
+| `LargeBodyResourceTest` | a 32 MiB streaming download consumes every byte without buffering the payload on the heap | structural + machine-dependent heap bound |
+| `PoolSaturationPerformanceTest` | 24 concurrent calls against a 4-connection pool all reach a terminal outcome; none hang | structural |
+| `Http2StreamSaturationTest` | 32 concurrent reactive streams share a 2-connection pool and complete | structural |
+| `OAuthRefreshContentionTest` | 100 genuinely concurrent callers produce exactly one token request | structural |
+| `RuntimeRotationDrainTest` | 50 rotations close all 50 retired generations and leave no drain thread | structural |
+
+## Recording a baseline
+
+When certifying a deployment, record alongside the numbers: the exact command, the commit, hardware,
+JVM flags, the profile YAML under test, p50/p95/p99/max, peak heap, peak direct memory, thread count,
+connection count, physical attempt count, and error count. A latency figure without its profile and
+hardware is not a baseline; it is an anecdote.
+
+| Field | Value |
+|---|---|
+| Command | _fill in at certification time_ |
+| Commit | _fill in_ |
+| Hardware / JVM | _fill in_ |
+| Profile under test | _fill in_ |
+| p50 / p95 / p99 / max | _fill in_ |
+| Peak heap / direct memory | _fill in_ |
+| Threads / connections | _fill in_ |
+| Physical attempts / errors | _fill in_ |
+
+The table is intentionally left unfilled in the repository: publishing numbers measured on a build
+agent as if they were a certified baseline would be worse than having none.
@@ -0,0 +1,47 @@
+# HTTP Client Platform — Release Checklist
+
+A release is complete when each item below is demonstrated by a command, not by review.
+
+## Gates
+
+```bash
+cd src
+./gradlew :adapter:outbound:httpclient:test --console=plain
+./gradlew :adapter:outbound:httpclient:httpClientStableContractTest --console=plain
+./gradlew :adapter:outbound:httpclient:httpClientSecurityTest --console=plain
+./gradlew :adapter:outbound:httpclient:httpClientBlockHoundTest --console=plain
+./gradlew :adapter:outbound:httpclient:spring62CompatibilityTest --console=plain
+./gradlew :adapter:outbound:httpclient:spring70CompatibilityTest --console=plain
+./gradlew :adapter:outbound:httpclient:httpClientFailureInjectionTest --console=plain   # needs Docker
+./gradlew :adapter:outbound:httpclient:httpClientPerformanceTest \
+  -Pperformance.assertions.enabled=true --console=plain
+./gradlew verifyCleanArchitectureDependencies --console=plain
+./gradlew :app-bootstrap:test --tests '*CleanArchitectureTest' --console=plain
+python3 ../scripts/verify-httpclient-docs.py
+```
+
+## Completion criteria (design §33)
+
+- [ ] Typed clients are the default entry point; H2 and H3 are separately authorised.
+- [ ] H1–H4 cannot bypass timeout, host, TLS, auth, size, or observation policy.
+- [ ] Apache, JDK, and Reactor produce identical result and exception metadata.
+- [ ] Pool, DNS, connect, TLS, and retry backoff all fit inside the effective deadline.
+- [ ] Every extra attempt is explained by idempotency, replayability, evidence, deadline, and budget.
+- [ ] Non-idempotent `SENT_NO_RESPONSE` surfaces as `HttpAmbiguousExecutionException`.
+- [ ] Pool and buffers are reclaimed after unread bodies, decode errors, cancels, and size rejections.
+- [ ] OAuth2 refresh is single-flight and 401 replay happens at most once.
+- [ ] Trust-all and hostname-verification bypass fail at startup.
+- [ ] Canonicalisation, DNS/IP validation, redirect revalidation, and egress control all pass.
+- [ ] No transparent retry occurs after the first delivered byte.
+- [ ] No platform code blocks a Reactor event loop, proven by a BlockHound self-check.
+- [ ] The negotiated wire protocol matches what the support matrix claims per transport.
+- [ ] Logical calls and attempts are separate metrics with no forbidden label.
+- [ ] DNS, pool, TLS, reset, partial response, and HTTP/2 GOAWAY are reproducible.
+- [ ] Thread, heap, direct memory, pool, and retry budget bounds hold.
+- [ ] The support matrix, configuration reference, security guide, runbook, and migration guide match the code.
+
+## Experimental
+
+Jetty HTTP/3 stays Experimental until `Http3CapabilityReport` reports QUIC and TLS 1.3 and the
+contract subset it declares passes in a dedicated environment. It is never auto-configured by the
+Stable starter.
@@ -0,0 +1,87 @@
+# HTTP Client Platform — Repository Adaptation Contract
+
+**Design source:** `docs/superpowers/specs/2026-08-08-httpclient-platform-design.md`
+**Plan source:** `docs/superpowers/plans/2026-08-08-httpclient-platform-implementation-plan.md`
+
+The design package states its own adaptation rule:
+
+> 실제 Backend Skeleton 저장소가 제공되지 않았으므로 package 경로와 Gradle 구조는 설계서의 명시적
+> 구현 가정이다. 구현 전 저장소의 기존 convention과 root package에 맞춰 경로만 조정하고 공개 계약과
+> 정책 의미론은 유지한다.
+
+This file is the single record of *how* the design's assumed layout was mapped onto this repository.
+Only paths, build DSL, and composition-root ownership changed. Public contracts, policy order, and
+error semantics are implemented exactly as specified.
+
+## 1. Why the module layout differs
+
+The design assumes a greenfield library with 19 Gradle projects under `modules/httpclient/`.
+This repository is a Clean Architecture template whose **fail-closed registry**
+(`src/config/architecture/modules.json`, enforced by `src/settings.gradle` and
+`verifyCleanArchitectureDependencies`) declares **exactly 19 leaf identities**. Creating 19 more
+Gradle projects would violate HARD-STOP #5 in `AGENTS.md`.
+
+Therefore the design's 19 library modules become **package boundaries inside the registered leaf**
+`:adapter:outbound:httpclient`, with two exceptions driven by this repository's own rules:
+
+| Design module | Repository home | Reason |
+|---|---|---|
+| `httpclient-spring-boot-starter` | `:app-bootstrap` (`dev.caskeleton.bootstrap.autoconfigure.httpclient`) | This repository's composition root owns wiring and canonical activation; an adapter leaf must not auto-configure itself. |
+| `httpclient-testkit` | `:adapter:outbound:httpclient` `src/testkit/java/**/testkit` | The design forbids production modules depending on the testkit; a source set whose dependencies are declared only on the test configurations gives the same guarantee without a new Gradle project. It is its own source set rather than part of `test` because three lanes consume it — `test`, `httpClientPerformanceTest` and `jmh` — and reaching into `sourceSets.test.output` from `jmh` compiled under Gradle but could not be modelled by an IDE, which classifies a source set as test source only when a `Test` task runs its output and forbids main source from reading test source. `PlatformClasses` excludes the source set's output so the boundary rules keep meaning production classes. |
+
+The package boundary is enforced by ArchUnit rules (`PublicApiArchitectureTest`,
+`HttpClientModuleBoundaryTest`) that reproduce the design's module dependency table.
+
+## 2. Package mapping
+
+Root package: `io.backend.skeleton.httpclient` → `dev.caskeleton.adapter.outbound.httpclient`.
+
+| Design module | Design package | Repository package |
+|---|---|---|
+| `httpclient-core-api` | `…httpclient.api` (+ `.body`, `.error`, `.operation`, `.result`) | `dev.caskeleton.adapter.outbound.httpclient.api` (+ same subpackages) |
+| `httpclient-profile` | `…httpclient.profile` | `…outbound.httpclient.profile` |
+| `httpclient-transport-spi` | `…httpclient.transport` | `…outbound.httpclient.transport` |
+| `httpclient-transport-apache` | `…httpclient.apache` | `…outbound.httpclient.apache` |
+| `httpclient-transport-jdk` | `…httpclient.jdk` | `…outbound.httpclient.jdk` |
+| `httpclient-restclient` | `…httpclient.restclient` | `…outbound.httpclient.restclient` |
+| `httpclient-resilience` | `…httpclient.resilience` | `…outbound.httpclient.resilience` |
+| `httpclient-auth` | `…httpclient.auth` | `…outbound.httpclient.auth` |
+| `httpclient-security` | `…httpclient.security` | `…outbound.httpclient.security` |
+| `httpclient-observability` | `…httpclient.observation` | `…outbound.httpclient.observation` |
+| `httpclient-transport-reactor-netty` | `…httpclient.reactor` | `…outbound.httpclient.reactor` |
+| `httpclient-webclient` | `…httpclient.webclient` | `…outbound.httpclient.webclient` |
+| `httpclient-service-client` | `…httpclient.service` | `…outbound.httpclient.service` |
+| `httpclient-dynamic-target` | `…httpclient.dynamic` | `…outbound.httpclient.dynamic` |
+| `httpclient-resttemplate-migration` | `…httpclient.migration` | `…outbound.httpclient.migration` |
+| `httpclient-spring7-service-groups` | `…httpclient.spring7` | `…outbound.httpclient.spring7` |
+| `httpclient-jetty-http3-experimental` | `…httpclient.http3` | `…outbound.httpclient.http3` |
+| `httpclient-spring-boot-starter` | `…httpclient.autoconfigure` | `dev.caskeleton.bootstrap.autoconfigure.httpclient` |
+| `httpclient-testkit` | `…httpclient.testkit` | `…outbound.httpclient.testkit` (`testkit` source set) |
+
+## 3. Other deliberate substitutions
+
+| Design assumption | Repository reality | Adaptation |
+|---|---|---|
+| Gradle Kotlin DSL, `build-logic` convention plugin | Groovy DSL, root `build.gradle` conventions, `LockMode.STRICT` dependency locking | Dependencies declared in `src/adapter/outbound/httpclient/build.gradle`; `gradle.lockfile` regenerated. |
+| Spring Framework 6.2 baseline with 7.0 compatibility | Spring Boot 4.0.8 / Spring Framework 7.0 is the repository baseline | Common code targets the Spring 6.2 **API surface** (no 6.2-only or 7.0-only classes in common packages). The Spring 7 HTTP Service Group integration stays isolated in `…httpclient.spring7`, exactly as the design requires. |
+| `settings.gradle.kts` module registration | Fail-closed registry | No registry change; leaf identity, gradle path, allowed dependencies unchanged. |
+| Design §6.2 grades Apache HttpClient 5 as HTTP/2-capable | Spring's blocking factory drives Apache's **classic** client, which is HTTP/1.1 only; HTTP/2 lives in Apache's async client | `ApacheBlockingTransportProvider` declares HTTP/1.1 and rejects an HTTP/2 profile at startup. Blocking HTTP/2 is served by the JDK transport, measured by `NegotiatedProtocolContractTest`. |
+| Design §28.1 names WireMock for stateful fixtures | WireMock's Jetty modules bind a different Jetty 12 ABI than the Boot-managed one this module already needs for HTTP/3, and fail at server start | `StatefulUpstream` provides path-keyed stateful responses on the existing fixture server; the WireMock dependency was removed rather than worked around with a shaded jar |
+| Per-task `git commit` | `AGENTS.md`: commit policy is `human-only` | Implementation is delivered unstaged; commits are the human's action. This is the only plan step intentionally not executed, and it is recorded here. |
+| `docs/httpclient/**`, `.github/workflows/httpclient-*.yml`, `scripts/verify-httpclient-docs.py` | Repository already owns `docs/` and `.github/workflows/` | Created at the same repository-relative paths. |
+
+## 4. What is unchanged from the design
+
+- H1 / H2 / H3 / H4 exposure rules and the forbidden native-engine signatures.
+- `ExecutionEvidence`, `BodyReplayability`, `OperationIdempotency`, `AttemptStage`, `FailureCategory`.
+- `HttpOperation`, `HttpCallResult`, `BodySource`, `ResponseType`, `BlockingStreamingResponse`.
+- The complete stable exception hierarchy and `HttpFailureMetadata` redaction rules.
+- Named Client Profile schema, startup validation codes, and operation override direction.
+- Effective deadline formula, attempt budget, and streaming setup/idle split.
+- Retry eligibility inputs, the ordered decision table, retry budget, and backoff rules.
+- Circuit → Rate Limiter → Bulkhead attempt order and logical admission placement.
+- OAuth2 cache key, single-flight refresh, and the 401 replay-at-most-once rule.
+- TLS allow/forbid lists and permanent-failure classification.
+- Dynamic Target canonicalization → all-answer DNS validation → pinning → redirect revalidation.
+- Low-cardinality tag allowlist, forbidden labels, trace and logging rules.
+- Runtime generation swap and drain semantics.
@@ -0,0 +1,71 @@
+# Retry and Ambiguity
+
+The platform never decides a retry from the HTTP method alone (design D-09). A second attempt
+happens only when idempotency, body replayability, execution evidence, deadline, and retry budget
+all permit it.
+
+## Execution evidence
+
+| Evidence | Meaning | Typical cause |
+|---|---|---|
+| `NOT_SENT` | Proven that the server never received the request | profile rejection, pool timeout, DNS failure, connect failure, pre-request TLS failure, HTTP/2 `REFUSED_STREAM` |
+| `SENT_NO_RESPONSE` | Some or all of the request was written, no final header arrived | partial write, response-header timeout, connection reset |
+| `RESPONSE_RECEIVED` | Final headers arrived, whatever the status | 2xx, 4xx, 5xx, redirect |
+| `PARTIAL_RESPONSE` | Headers and part of the body arrived | reset during decode, interrupted stream |
+
+`NOT_SENT` is only produced by a stage failure that proves it. A generic engine I/O error is never
+upgraded to `NOT_SENT`, because that is exactly how a timeout becomes a duplicate payment.
+
+## Body replayability
+
+| Body | Replayability |
+|---|---|
+| immutable `byte[]` | `REPLAYABLE` |
+| DTO plus a deterministic codec | `REPLAYABLE` |
+| reopenable file or resource supplier | `REOPENABLE` |
+| a single `InputStream` instance | `ONE_SHOT` |
+| publisher factory | as declared |
+| publisher instance | `ONE_SHOT` |
+| multipart | the weakest part |
+
+## Decision order
+
+`DefaultRetryEligibilityEngine` evaluates in this order, and a later rule can never re-enable
+something an earlier one forbade:
+
+1. attempts exhausted → `RetryDenied.maxAttempts()`
+2. retry budget empty → `RetryDenied.budgetExhausted()`
+3. body not replayable → `RetryDenied.bodyNotReplayable()`
+4. first byte already delivered → `RetryDenied.responseAlreadyDelivered()`
+5. runtime draining → `RetryDenied.runtimeDraining()`
+6. remaining deadline below the minimum attempt budget → `RetryDenied.deadline()`
+7. permanent failure category → `RetryDenied.permanentFailure(...)`
+8. `SENT_NO_RESPONSE` on an operation that is not safely idempotent → `AmbiguousFailure`
+9. status- and failure-specific rules
+
+## Status rules
+
+| Status | Decision |
+|---|---|
+| 408 | retry inside deadline and budget |
+| 425 | at most one retry, first attempt only |
+| 429 | retry inside `Retry-After`, deadline, and budget |
+| 401 | one refresh-and-replay, safe replayable operations only |
+| 500 | denied unless the upstream registered it as transient **and** the operation is safely idempotent |
+| 502, 503, 504 | retry for safely idempotent operations; ambiguous otherwise |
+| other 4xx | denied |
+
+## Ambiguity
+
+A non-idempotent request that reached `SENT_NO_RESPONSE` raises
+`HttpAmbiguousExecutionException`. It is a third answer on purpose: retrying may duplicate a side
+effect, and reporting a plain failure would tell the caller the request did not happen, which may
+be false. The caller reconciles, usually by querying the upstream or replaying with an idempotency
+key.
+
+## Budget and backoff
+
+Retry tokens come from a per-upstream token bucket sized as a fraction of real traffic, so a failing
+upstream cannot be flooded by retries from a healthy fleet. Backoff is exponential with full or
+decorrelated jitter, bounded by `max-backoff`, by `Retry-After`, and by the remaining deadline. No
+connection and no bulkhead permit is held while a backoff is waiting.
@@ -0,0 +1,75 @@
+# HTTP Client Platform — Security Guide
+
+## What the platform owns
+
+`Authorization`, `Proxy-Authorization`, `Host`, `Content-Length`, `Transfer-Encoding`,
+`Traceparent`, `Tracestate`, `Baggage`, and (unless a profile opts in) `Cookie` are platform-owned.
+A caller cannot set them. `Idempotency-Key` is accepted only when the operation declares it. Any
+header name or value containing CR or LF is rejected before the request is built.
+
+## Target policy
+
+A trusted profile accepts only a profile-relative URI template. An absolute URI is rejected rather
+than sanitised: varying the destination is what H3 is for, and H3 has its own policy, credentials,
+and address validation. Template variables are encoded per component, so a value containing `/`,
+`?`, or `#` cannot change the shape of the request.
+
+## TLS
+
+Allowed: TLS 1.2 and 1.3, hostname verification, the JVM trust store, a per-profile custom CA, a
+per-profile client certificate, mTLS, SNI and ALPN, and certificate rotation through a new runtime
+generation.
+
+Forbidden and unrepresentable: a trust-all trust manager, disabled hostname verification, ignoring
+certificate errors, automatically trusting a production self-signed certificate, falling back to
+plaintext after an HTTPS failure, and writing key material into configuration or logs.
+
+Unknown CA, hostname mismatch, expired certificate, revoked certificate, protocol mismatch, and a
+missing client certificate are permanent. Only a transient handshake timeout may be retried, inside
+the deadline.
+
+## Dynamic Target (SSRF)
+
+Every hop — the first one included — runs the whole flow:
+
+1. strict URI parse
+2. scheme allowlist
+3. reject userinfo and invalid ports
+4. IDNA-canonicalise the host
+5. host allowlist or suffix policy
+6. resolve **every** A and AAAA answer
+7. normalise each address, including IPv4-mapped IPv6
+8. reject loopback, link-local, RFC1918, ULA, carrier-grade NAT, unspecified, multicast, cloud
+   metadata, and organisation-defined ranges
+9. pin the connection to the approved addresses through the same validated resolver
+10. apply response size and content policy
+11. repeat for each redirect
+
+Any forbidden address in the answer set rejects the whole target. Validating only the first answer
+would let a host that resolves to one public and one private address through.
+
+Dynamic profiles inherit no API key, OAuth token, Cookie, or default header, and no Cookie jar is
+created. A specific host may be granted a credential only through an explicitly registered
+`DynamicCredentialBinding`.
+
+Application-level validation is not sufficient on its own. A network control — Kubernetes
+NetworkPolicy, service-mesh egress policy, firewall, or proxy ACL — is an operational completion
+requirement.
+
+## Redirects
+
+Disabled by default. Engine redirect handling is off in every transport so the platform can
+re-validate each hop. 307 and 308 preserve method and body and are therefore allowed only for a
+replayable body. Cross-origin hops are refused unless the profile opts in, and when they are
+allowed `Authorization`, `Proxy-Authorization`, `Cookie`, and API-key headers are stripped.
+
+## Observability
+
+Allowed tags: `clientName`, `operationName`, `method`, `uriTemplate`, `status`, `outcome`,
+`transport`, `protocol`, `timeoutType`, `retryReason`, `evidence`, `circuitState`.
+
+Rejected outright: full URL, query parameters, path variable values, user ID, raw tenant ID,
+resolved IP, API key, token, Cookie, idempotency key, request or response body, exception message.
+
+Failures are logged once, structured, at the end of a logical call. Retry attempts are DEBUG or span
+events. URLs appear only as templates.
@@ -0,0 +1,52 @@
+# Streaming and Large Bodies
+
+## Response lifecycle
+
+A blocking streaming download returns `BlockingStreamingResponse`, never a bare `InputStream`.
+Closing is idempotent and always releases the connection — after a full read, a partial read, a
+decode failure, or a size rejection. The status is validated before any body byte is delivered, so a
+failed download never becomes a half-consumed stream the caller has to reason about.
+
+A reactive download emits bounded `DataBuffer` values. Buffers are released on completion, error,
+and cancellation; a dropped buffer is direct memory nobody returns.
+
+Wire bytes and decoded bytes are bounded independently, because a compressed payload passes a wire
+check and then expands. Limits are enforced while reading, not after buffering.
+
+## The first-byte boundary
+
+```text
+response headers received
+  → nothing delivered yet
+     → a read-only operation may still be retried
+  → first InputStream read or first Flux onNext
+     → transparent retry is permanently disabled
+```
+
+`FirstByteDeliveryGuard` latches once and never resets. Retrying after delivery would replay a
+stream the caller has already partly consumed, producing duplicated or reordered data that no
+downstream code can detect.
+
+## Request bodies
+
+A reopenable body is opened once per attempt, which is what makes it replayable; reusing the
+previous stream would silently send an empty body on the retry. A one-shot stream or publisher
+instance is never retried. `ReactiveBodySource` takes a publisher *factory* rather than a publisher
+so a reactive body can honestly declare itself replayable.
+
+A multipart body is exactly as replayable as its weakest part.
+
+## Server-sent events
+
+Three budgets stay separate:
+
+- `setupDeadline` — establishing the stream
+- `streamingIdleTimeout` — silence once it is open
+- `maxStreamDuration` — optional total lifetime
+
+Applying the request-shaped `total-call` timeout to an SSE subscription would terminate a perfectly
+healthy stream on schedule, so it is not applied.
+
+`Last-Event-ID` is opt-in. Replaying from an id is only correct when the producer guarantees it;
+sending it blindly can skip or duplicate events. Reconnects consume the retry budget like any other
+physical attempt, and cancelling the subscription stops both the stream and any pending reconnect.
@@ -0,0 +1,88 @@
+# HTTP Client Platform — Support Matrix
+
+Grades follow design §6 and §29. A row is **Stable** only when the cross-transport contract suite
+proves it; anything the suite cannot prove is **Experimental** and says so.
+
+## Spring API
+
+| API | Grade | Role | Constraint |
+|---|---|---|---|
+| `RestClient` | Stable | Blocking execution | Bounded concurrency and an effective deadline are mandatory |
+| `WebClient` | Stable | Reactive, streaming, SSE | No blocking work on the event loop |
+| HTTP Service Client (`@HttpExchange`) | Default | Declarative typed client | Operation metadata is mandatory |
+| `RestTemplate` | Migration only | Moving existing calls | No new profile or feature |
+| Generic Exchange (H2) | Restricted | Dynamic method, path, body | Base URL and policy are immutable |
+| Dynamic Target (H3) | Restricted | User-supplied URL | Separate SSRF policy; inherits no credential |
+| Native engine | Internal | Engine-specific configuration | Never an application-facing API |
+
+## Transports
+
+| Transport | Blocking | Reactive | HTTP/1.1 | HTTP/2 | HTTP/3 | Grade | Verified by |
+|---|---:|---:|---:|---:|---:|---|---|
+| Apache HttpClient 5 (classic) | yes | no | yes | **no** | no | Stable (blocking default) | `httpClientStableContractTest`, `NegotiatedProtocolContractTest` |
+| JDK HttpClient | yes | `sendAsync` | yes | yes (TLS/ALPN) | no | Stable (lightweight, blocking HTTP/2) | `NegotiatedProtocolContractTest` |
+| Reactor Netty | limited | yes | yes | yes | experimental | Stable (reactive default) | `NegotiatedProtocolContractTest` |
+| Jetty | facade | yes | yes | yes | yes | **Experimental** | `Http3OptInTest` only |
+| Simple request factory | yes | no | limited | no | no | Local test only | rejected in production by `ClientProfileValidator` |
+
+### Apache is HTTP/1.1 here, and why
+
+Design §6.2 grades Apache HttpClient 5 as HTTP/2-capable, and the library is — in its **async**
+client. Spring's `HttpComponentsClientHttpRequestFactory` drives the **classic** client, which
+speaks HTTP/1.1 only. `NegotiatedProtocolContractTest` measures this rather than assuming it: the
+classic client fails outright against a prior-knowledge h2c server.
+
+So `ApacheBlockingTransportProvider.capabilities()` declares HTTP/1.1, and a profile that pairs
+Apache with `HTTP_2` is rejected at startup instead of quietly running HTTP/1.1 while this table
+claims otherwise. **Blocking HTTP/2 is served by the JDK transport**; reactive HTTP/2 by Reactor
+Netty. Both are measured from the client after a real TLS handshake, not read from configuration.
+
+The JDK transport declares `routeScopedPool=false`, `boundedPendingAcquireQueue=false`, and
+`dynamicTargetStable=false`. A profile that needs any of those is rejected at startup rather than
+served with weaker guarantees. Choosing between Apache and JDK is therefore a real trade: Apache
+gives route-scoped pooling and Dynamic Target pinning, JDK gives HTTP/2.
+
+## Capability gates
+
+| Capability | Gate |
+|---|---|
+| Dynamic Target (H3) | Apache and Reactor Netty only; JDK and Jetty are rejected |
+| HTTP/3 | `experimentalAcknowledgement` must equal `I_ACCEPT_HTTP3_EXPERIMENTAL_SEMANTICS` |
+| Cross-origin redirect | opt-in per profile; credentials are stripped on the hop |
+| Retry | evidence-based; never enabled by HTTP method alone |
+
+## CI matrix
+
+| Profile | Frequency | Release gate | Task |
+|---|---|---|---|
+| Spring Framework 7.0 (repository baseline) | every PR | required | `spring70CompatibilityTest` |
+| Spring Framework 6.2 API surface | every PR | required | `spring62CompatibilityTest` |
+| Apache HC5 + RestClient | every PR | required | `httpClientStableContractTest -Phttpclient.contract.transports=apache` |
+| JDK HttpClient + RestClient | every PR | required | `httpClientStableContractTest -Phttpclient.contract.transports=jdk` |
+| Reactor Netty + WebClient | every PR | required | `httpClientStableContractTest -Phttpclient.contract.transports=reactor` |
+| SSRF / cardinality suite | every PR | required | `httpClientSecurityTest` |
+| Toxiproxy fault suite | nightly, release | required | `httpClientFailureInjectionTest` |
+| Event-loop blocking (BlockHound) | every PR | required | `httpClientBlockHoundTest` |
+| Performance certification | nightly, release | required | `httpClientPerformanceTest -Pperformance.assertions.enabled=true` |
+| Jetty HTTP/3 | nightly | Experimental, non-blocking | `test -Phttp3.tests.enabled=true` |
+
+### Known limitation of the Spring 6.2 lane
+
+This repository's Spring Boot 4.0 baseline pins Spring Framework 7, so a real 6.2 runtime cannot be
+resolved here. `spring62CompatibilityTest` therefore verifies the **API surface**: the common
+packages must not reference any Spring 7-only type, and `org.springframework.web.service.registry`
+is confined to `…httpclient.spring7`. Executing the suite against an actual 6.2 distribution
+requires a host project on that line. This limitation is stated rather than hidden behind a passing
+check.
+
+
+## What the suites do not prove
+
+Stated so the matrix is read as a measurement rather than an aspiration.
+
+| Gap | Why | What is proven instead |
+|---|---|---|
+| HTTP/2 frame injection (`REFUSED_STREAM`, arbitrary `GOAWAY`) | The fixture server exposes no frame-level control, and a purpose-built h2 server is a larger dependency than the guarantee is worth here | `Http2EvidenceMapperTest` proves the frame → evidence mapping, and `NegotiatedProtocolContractTest` proves h2 is really negotiated |
+| Netty buffer-leak detection | Netty reports a leak when an unreferenced buffer is collected, which the suite does not force | `NettyLeakDetectionExtension` asserts the PARANOID detector is live and reports nothing; explicit release assertions in the streaming suites are the primary guarantee |
+| Spring 6.2 runtime | This repository's Boot 4.0 baseline pins Spring 7 | `spring62CompatibilityTest` confines the common packages to the 6.2 API surface |
+| Performance latency baseline | Numbers measured on a build agent are not a certification | `httpClientPerformanceTest` asserts structural bounds unconditionally; latency and heap bounds run under `-Pperformance.assertions.enabled=true` |
@@ -0,0 +1,64 @@
+# Entity Mapping Guide
+
+Design §10-§13. The rules here exist because each one has a failure mode that is invisible in review
+and expensive in production.
+
+## The domain owns the model
+
+The platform defines no business entity. Table names, column semantics, keys, unique and check
+requirements, associations, cascade rules, lock policy, and soft-delete policy all belong to the
+domain module. There is no `GenericRepository<T, ID>` and no platform base repository, because a
+single generic API forces every aggregate through the same operations — and one aggregate's later
+requirement then changes behaviour for all of them.
+
+## Entities must be proxyable
+
+- Not `final`. Hibernate creates a lazy proxy by generating a subclass; a final entity cannot be
+  subclassed, so *every* association to it loads eagerly whatever the mapping says. Nothing errors.
+- A non-private no-arg constructor. The provider instantiates entities reflectively before
+  populating fields.
+
+`EntityMappingCondition` in the testkit enforces both.
+
+## Identifiers
+
+Default to a sequence with an `allocationSize` that matches the migration's `INCREMENT BY`. When
+they disagree, the provider hands out identifiers the sequence has not reserved and the collision
+surfaces later as a primary-key violation under load.
+
+`GenerationType.IDENTITY` is supported and limited: the key is assigned on insert, so the provider
+must execute each insert immediately to learn it, which disables JDBC insert batching entirely.
+`HibernateBatchConfigurationGuard` fails a batch profile that targets an IDENTITY entity rather than
+letting the import silently run an order of magnitude slower.
+
+UUIDv7 (`UuidV7Generator`) is the application-side option. It is preferred over UUIDv4 for a primary
+key because v4 is uniformly random: every insert lands on a random leaf of the B-tree, so the index
+never stays in cache and write amplification grows with the table.
+
+## Values
+
+- Enums are `EnumType.STRING` or an explicit converter. **Never** `ORDINAL` — it stores the
+  constant's position, so inserting a new constant anywhere but the end silently reinterprets every
+  existing row.
+- Money is `BigDecimal` with explicit precision and scale. `double` cannot represent `0.1`, so sums
+  drift and reconciliation disagrees with the ledger.
+- `Duration` goes through a converter that stores milliseconds. The ISO-8601 text form sorts and
+  compares wrongly in SQL.
+- `Instant` and `OffsetDateTime` map differently; a column typed for one cannot faithfully store the
+  other.
+
+## Associations
+
+- To-one associations are `LAZY`. JPA's default is `EAGER`, which means every query that loads a
+  child also queries for its parent — the most common accidental N+1 in a JPA application.
+- The owning side holds the foreign key. Adding to the inverse collection alone leaves the row
+  unlinked, so aggregates expose an association helper that sets both sides.
+- `CascadeType.ALL` with `orphanRemoval` is correct only for a child the aggregate genuinely owns.
+  Between independent aggregates it deletes rows another part of the system still owns.
+
+## Entities never leave the transaction
+
+A controller must not return an entity, or a collection or `Optional` of one. Response serialisation
+happens after the transaction closes, so a lazy association touched by the serialiser either throws
+or — with OSIV on, which this platform forbids — issues a query from the view layer, one per element.
+`EntityExposureCondition` checks generic type arguments, not just the erased return type.
@@ -0,0 +1,43 @@
+# Experimental Promotion Checklist
+
+`ExperimentalPromotionGate` evaluates this checklist. Every technical item, then the ADR.
+
+## Technical evidence
+
+- [ ] **Compatibility** — the Stable contract suite passes on the experimental target, twice, on two
+      supported patch releases. One passing run is a coincidence.
+- [ ] **Security** — for tenancy features, cross-tenant read *and* write are both proven impossible,
+      including through native SQL, bulk DML, `getReference`, and the second-level cache. A filter
+      that covers only entity queries covers none of those.
+- [ ] **Failure** — connection reuse does not leak tenant context; a failover does not silently route
+      a read-after-write to a stale replica; the commit-ambiguity scenarios still behave.
+- [ ] **Migration** — per-tenant migration is resumable after a partial failure, and rate-limited.
+      With one schema per tenant, a run is N independent migrations and "it failed" is not an answer.
+- [ ] **Performance** — pool capacity, replica lag under load, and per-tenant memory are measured,
+      not estimated. Database-per-tenant fails as a sum, not as an individual pool.
+
+## Decision
+
+- [ ] **Reviewed ADR** — recording what is being promised, the operational burden it carries, and
+      what would cause it to be withdrawn.
+
+The ADR is not a formality. The technical suites establish that something works; the ADR records
+that the platform should promise it, which is a different question with a different cost.
+
+## What does not count as evidence
+
+- The version being generally available.
+- The feature working in one environment.
+- A passing suite that skipped because Docker was unavailable.
+- A green lane whose assertions were relaxed to make it pass.
+
+## Outcomes
+
+| Decision | Meaning |
+|---|---|
+| `BLOCKED_TECHNICAL` | at least one suite has not passed |
+| `BLOCKED_MISSING_ADR` | evidence is complete; no reviewed decision exists |
+| `ELIGIBLE_FOR_STABLE_REVIEW` | both; Stable review may begin |
+
+The two blocked states are distinct because they need different work: one needs evidence, the other
+needs a decision.
@@ -0,0 +1,43 @@
+# Experimental Support Matrix
+
+Everything here is off unless its `backend.jpa.experimental.*` flag is explicitly true, and none of
+it is part of the Stable composition.
+
+| Feature | Flag | State |
+|---|---|---|
+| Shared-schema multi-tenancy (column) | `backend.jpa.experimental.multitenancy-column` | Experimental |
+| PostgreSQL RLS multi-tenancy | `backend.jpa.experimental.multitenancy-rls` | Experimental |
+| Schema-per-tenant | `backend.jpa.experimental.multitenancy-schema` | Experimental |
+| Database-per-tenant | `backend.jpa.experimental.multitenancy-database` | Experimental |
+| Consistency-aware read replica | `backend.jpa.experimental.read-replica` | Experimental |
+| Jakarta Persistence 4.0 lane | `backend.jpa.experimental.jakarta-persistence-4` | Experimental |
+| Hibernate ORM 8 lane | `backend.jpa.experimental.hibernate-8` | Experimental |
+| PostgreSQL 19 lane | `backend.jpa.experimental.postgresql-19` | Experimental |
+
+Presence on the classpath is not consent. `ExperimentalFeatureGate` fails startup when a module is
+present and its flag is not set, because an experimental module can arrive transitively and a
+tenant-isolation feature that switched itself on would be the worst possible default.
+
+## Known constraints
+
+- Tenant context is fail-closed. An unbound tenant in a shared-schema deployment means a query with
+  no tenant predicate, which returns every tenant's rows.
+- A Hibernate filter is not the security boundary. It does not apply to native SQL, bulk DML,
+  `getReference`, or the second-level cache.
+- RLS requires all three of: `ENABLE ROW LEVEL SECURITY`, `FORCE ROW LEVEL SECURITY` (the owner is
+  otherwise exempt from its own policies), and a runtime role without `BYPASSRLS`.
+- Tenant bindings are transaction-local. A session-local setting survives the connection's return to
+  the pool.
+- `readOnly=true` never routes to a replica on its own. Read-after-write uses a consistency token or
+  the primary.
+- Unavailable replica lag evidence means the primary. Absence of evidence is not evidence of
+  freshness.
+- Per-tenant pools are bounded globally. Fifty tenants with a modest pool each is five hundred
+  connections against a server that permits a hundred.
+- Tenant ids never become metric tags. Tenant cardinality is unbounded by definition.
+
+## Lanes never change Stable
+
+A compatibility lane publishes nothing and changes no Stable contract. If Hibernate 8 generates
+different SQL for the fetch-pagination gate, that is a finding about Hibernate 8 — the 7.x gate keeps
+asserting what 7.x must do, because that is what deployments run.
@@ -0,0 +1,64 @@
+# Migration Guide
+
+Design §31-§32. Flyway owns the schema; Hibernate only validates.
+
+## Who may change the schema
+
+| Environment | Mode |
+|---|---|
+| local, test, dev | migrate at startup with the migration credential |
+| staging, prod | deployment-owned migration; the application validates only |
+
+Migrating from inside the application in production means every instance of a rolling deploy races
+to apply the same script, and the loser's failure is indistinguishable from a real one.
+
+`ddl-auto` is `validate` or `none`. Never `update`: it never drops or narrows anything, so it
+produces a schema that is neither the old one nor the one the migrations describe — silently, on
+whichever instance started first.
+
+## Validation fails closed and never repairs
+
+`FlywayValidationGate` throws `SchemaMismatchException` on a checksum mismatch, a missing migration,
+or a schema Hibernate disagrees with. It never calls `repair`.
+
+Repair rewrites the schema history table to match whatever scripts are on disk. That resolves the
+symptom by deleting the evidence: a checksum mismatch means the deployed script differs from the
+applied one, and the interesting question is which change is missing from this database. Repair
+makes that question unaskable. It exists only as an explicit admin operation with an operator, a
+reason, and an approval (design §8.4).
+
+Only Flyway's structured error codes reach the exception. Its messages embed the script path and
+part of the failing statement.
+
+## Concurrent index builds
+
+`CREATE INDEX CONCURRENTLY` cannot run inside a transaction block, and Flyway wraps migrations in
+one by default. The migration therefore needs a companion configuration:
+
+```conf
+# V42__order_index.sql.conf
+executeInTransaction=false
+```
+
+`ConcurrentIndexMigrationInspector` fails validation without it, and additionally requires the
+migration to contain nothing else. A failed concurrent build leaves an invalid index behind;
+recovering is a single `DROP INDEX` when the migration did nothing else, and a manual reconstruction
+of partial state when it did.
+
+An invalid index is not merely useless — the planner ignores it while every write still maintains
+it. `FailedConcurrentIndexRecovery` reports them with the statement to run, and deliberately does
+not drop them: an invalid index can also mean a build is still running, and the two are
+indistinguishable from the catalog alone.
+
+## Upgrade scenarios
+
+Three, each catching something the others do not:
+
+| Scenario | Catches |
+|---|---|
+| `empty` | an early migration edited to match a later one, no longer applying to a fresh database |
+| `previous-release` | the actual deployment path; the only one exercising this release's migrations |
+| `oldest-supported` | a migration that silently assumes state only recent databases have |
+
+Each asserts a data invariant, not just the schema version. A migration that renames a column and
+loses its contents leaves the version correct and the data gone.
@@ -0,0 +1,61 @@
+# Observability
+
+Design §37. What is measured, and what must never appear in a measurement.
+
+## Bounded tags, always
+
+Every JPA metric carries exactly five tags: persistence unit, operation, query, outcome, failure
+category. All five are registered identifiers, validated by `LowCardinality` at construction rather
+than at the registry — so an unbounded value fails where it was introduced instead of surviving
+until a dashboard stops loading.
+
+Never a tag: entity id, tenant id, SQL parameter, exception message, JDBC URL. Each is unbounded, so
+each creates a time series per row or per failure; several are also the data the platform keeps out
+of logs, which a metrics backend would store just as durably and export just as widely.
+
+## Transaction metrics
+
+| Meter | Why it exists |
+|---|---|
+| `jpa.transaction.duration` | the baseline |
+| `jpa.transaction.rollback` | rollback rate by failure category |
+| `jpa.transaction.timeout` | timeouts, distinct from other rollbacks |
+| `jpa.transaction.completion.unknown` | its own counter, deliberately |
+
+Completion-unknown gets a separate counter rather than being folded into failures. It is the one
+outcome that means a human has to look: every other failure is a transaction that definitely did not
+happen, while this one is a transaction that may have.
+
+## Query metrics
+
+`jpa.query.duration` and `jpa.query.rows`. Rows are measured as well as duration because a query
+that issues one statement and hydrates twenty thousand rows is fast per statement and catastrophic
+per request — a duration metric alone reports it as merely slow.
+
+## Retry metrics
+
+Attempts are metrics, not warnings. Optimistic conflicts and serialization failures are the expected
+cost of concurrency; logging each at WARN pages someone for a system working as designed, after
+which the retry log gets filtered out and takes the genuinely interesting entries with it.
+
+`jpa.retry.attempt`, `jpa.retry.attempts` (distribution per operation), `jpa.retry.exhausted`.
+
+## Query names in SQL
+
+`NamedStatementInspector` prefixes each statement with its registered query name as a SQL comment,
+which travels into `pg_stat_activity`, `auto_explain`, and the slow-query log. Without it, "which
+endpoint issues this query" is answered by grepping the codebase for fragments of SQL.
+
+## Diagnostics
+
+`SqlDiagnosticRedactor` removes string literals, numbers, and anything email-shaped before SQL
+reaches a log. Redaction is blunt on purpose: preserving "harmless" values would require knowing
+which columns hold personal data.
+
+## The actuator endpoint
+
+`jpaplatform` reports database major version, provider version, schema version, OSIV state, runtime
+role verification, and capability levels. It reports no JDBC URL, no username, no SQL, and no entity
+catalog — an actuator endpoint is reachable by anyone who reaches the management port, and each of
+those would be a free reconnaissance answer. It is read-only: an endpoint that could trigger a
+migration or a repair would be an admin capability exposed over HTTP.
--- a/Show More
+++ b/Show More