리뷰 두 건을 반영했다. 계약 - tech-log-tree.json 하나가 분해 계약이자 색인이다. 사람이 읽는 트리·Node Specification· 후보 대장은 없어졌고, 문서에 남아 있던 그 개념을 걷어냈다 - candidateScope — 후보를 찾는 SSOT 범위. 접어 넣은 제2부·제3부는 근거이지 후보가 아니다 - sourceRepository — 분석한 저장소의 경로·리비전·판단 근거. 리비전을 모르면 null 로 두고 지어내지 않는다. 갈래가 여럿이면 revisions - 검사기: 계약 미채택·PENDING·PROMOTE↔글감 양방향·candidateScope·sourceRepository 를 error/warn 으로 센다. 옛 스키마도 검사를 피하지 못한다. 테스트 22 → 31 기록 쓰기 - 템플릿 5종에 source·sourceRevision·topicName, Question 에 닫는 조건, 본문 없는 종류에서 assets 제거. 고정 절 개수 삭제 - check_evidence.mjs — 인용한 코드가 SSOT 에 있는지, 앵커가 SSOT 를 가리키는지, 제목이 계약과 같은지, 리비전이 저장소에 있는지. 게시된 기록에서 SSOT 와 다른 URL 을 잡았다 문체 - 문체 규칙의 정본을 ai-tells.md 로. explaining.md 의 질문체 제목·절 끝 대조 반복·그림 예고 규칙을 삭제해 충돌을 없앴다. 첫 절 「설명 뒤에 평가를 붙이지 않는다」에 지우는 사례 네 유형 - voice 스킬의 「독자 쪽을 본다」를 자료에 오독 기록이 있을 때로 좁히고, 평가만 더한 예시를 교체 - check_prose: 안내 문장을 요구하던 경고 제거, 문장이 끝나지 않은 채 문단이 끝나는 조각 검사 추가 Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
3.2 KiB
3.2 KiB
Tech Log Tree Decomposition Checklist
Selection
- Every analysis candidate carries a disposition, and
KEEP_IN_SSOTis used. dispositionReview: PENDINGis 0 — nothing reached the tree by recall alone.- Each
PROMOTEnode passes the independence test. - Candidates were discovered from
final/document.md, not fromanalysis/**. - Material found only in
analysis/**was added tofinal/document.mdfirst. candidateScopeis declared, and no candidate came from outside it.
Source integrity
- The tree records the source document hash, ledger hash, and project revision.
- Every node has at least one source anchor.
- Source anchors actually contain the material implied by the title.
- Runtime-dependent claims name evidence or use
NEEDS_EVIDENCE. readinessstates how well evidenced the node is. It does not state whether the record has been written or published.
Topic quality
- Every Topic has one reader question, and every node in it helps answer that question.
- No two Topics share a reader question.
- Topic is an engineering problem space rather than a directory name.
- A large Topic is split when its records no longer share useful relations/criteria.
Case
- There is a specific incident, experiment, failure, diagnosis, or verification sequence.
- The title can be understood without inventing a historical story.
- The conclusion is bounded by actual evidence.
- Observations that answer the same question with the same conclusion are one Case, as a table or sub-sections — not several partial Cases.
Concept
- It was added because a Case, Decision, or Question needs it, not by sweeping headings.
basis-versionnames what the explanation was written against.- The title names a mechanism, not an absence, a count, or an analysis-scope fact.
- No analysis section number or finding grade survives in the title.
Reference
- The rule is reusable beyond the originating incident.
- It is not the Case summary rewritten declaratively.
- Scope and exception can be stated from sources.
Open Question
- The answer is not already in the analysis.
- Known/unknown/next verification are separable.
- The design or a decision actually turns on the answer.
- Candidate options are included only when sources really considered them.
Decision
- A project choice is explicitly recorded or user-supplied.
technology is presentis not being treated as rationale.- The accepted cost is stated, not only the benefit.
NEEDS_DECISIONis used if the direction is only a recommendation.
Duplication and shape
- No two nodes have the same primary purpose.
- Relations are used instead of copying one record into another.
- Node count is bounded by what a reader would read, not by the analysis denominator.
Parity
- Every node traces back to a
PROMOTEcandidate, and everyPROMOTEcandidate has a node. countsmatches whatbuild-tech-log-tree.pyproduces.unlistedis empty — no record file exists outside the contract.python3 scripts/verify-tech-log-tree.py <project>reports 0 errors.