workflow-id: hyeonworks-vnext-v1-direction
task-id: review-visual-craft-r4
mode: divergent
tier: standard
model: sonnet
effort: high
assigned-lens: visual-craft
target-role-agent: DES-VISUAL
workspace: hyeonworks
target-repo: hyeonworks
objective: exact revision-4 winner ENG-FE-20260718T144700Z를 visual-craft lens로 독립 검토한다.
output-format: report.yaml + report-path와 1줄 bottom-line
allowed-tools: [Read, view_image, exec_command, apply_patch]
task-boundaries: >-
  visual-craft와 computed contrast만 평가하고 이전 review는 읽지 않는다. distinctiveness를 대신 판정하거나 prototype을 수정하지 않고
  DES-VISUAL-20260718T145005Z.report.yaml만 작성한다. exact target SHA는
  8b208279e12f9aea0fa6d7a3c320811b0e72ddb74c6afac64cddadf4b157cf76이다.
non-goals: [distinctiveness 재판정, 다른 lens 판정, prototype 수정, 외부 네트워크]
must-read:
  - hyeonworks/completion-records/hyeonworks-vnext-v1-direction/ENG-FE-20260718T144700Z.report.yaml
  - hyeonworks/design-direction/hyeonworks-vnext-v1/prototype/core-flow.yaml
  - hyeonworks/design-direction/hyeonworks-vnext-v1/prototype/dist/styles.css
  - hyeonworks/design-direction/hyeonworks-vnext-v1/prototype/scripts/verify_flow.cjs
  - hyeonworks/design-direction/hyeonworks-vnext-v1/prototype/preview.w360.png
  - hyeonworks/design-direction/hyeonworks-vnext-v1/prototype/preview.w768.png
  - hyeonworks/design-direction/hyeonworks-vnext-v1/prototype/preview.w1280.png
  - hyeonworks/design-direction/hyeonworks-vnext-v1/prototype/state-previews
inherited-decisions:
  - {decision-id: DES-DIRECTOR-20260718T115933Z, summary: "editorial evidence, AA reading, responsive craft"}
acceptance-tests:
  - {criterion: "actual backgrounds의 small-text/control 대비와 responsive finish 판정", verify: "python3 .claude/hooks/validate_report.py hyeonworks/completion-records/hyeonworks-vnext-v1-direction/DES-VISUAL-20260718T145005Z.report.yaml"}
evidence-plan:
  - {claim: "computed contrast and responsive finish", grade: E3, receipt: "browser metrics, CSS and exact state renders"}
expected-output:
  report-header: {bottom-line: '', decision-needed: {needed: false, approver: null}, confidence: {value: Med, derived-from: evidence}, risks: [], evidence: []}
token-budget: {max-input-tokens: 60000, max-output-tokens: 8000}
method-selection: {method-id: art-direction, candidates: [art-direction, compare-directions]}