사용 방법
- 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
- 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
- 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/sergekostenchuk/ui-ux-agent-skill-system --skill workflow-compliance-supervisor시작 전 확인
- 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
- 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 Skill 문서
원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.
Workflow Compliance Supervisor
This skill is the governance layer for journey-driven UI/UX projects. It checks whether the system did what it promised, whether skipped stages are justified, and whether the user sees honest reasons and next actions.
It does not execute design/build tasks and does not override senior-ui-ux-orchestrator.
Operating Modes
gate: check the current project before continuing or closing a phase.review: audit an existing task plan, progress state, progress screen, and evidence set.log: append or validateworkflow-log.jsonlevents.fixture-test: run lazy-skip and fake-evidence fixtures.explain: explain to the user why a stage is paused, skipped, returned to rework, or blocked.
Required Workflow
- Read references/decision-framework.md.
- Read references/evidence-and-log-rules.md.
- Read references/user-messaging.md if the output will be shown to a non-technical user.
- Locate the canonical project
TASK-PLAN.md. - Locate the declared journey registry. Default local registry:
$UIUX_SKILL_SYSTEM_ROOT/data/journey-registry.json. - Compare declared journey stages with task-plan claims, progress state, reports, screenshots, and logs.
- Produce one decision per affected stage:
continue,pause_for_user,return_to_rework,skip_with_reason, orblock. - For every skip, require an acceptable reason code, impact, and next action.
- For every
done/Ranclaim, require real evidence. - Write or update a sanitized report/log only when the user asked for execution or validation.
Decision States
continue: evidence is sufficient or deviation has no material impact.pause_for_user: missing user data, approval, budget, credentials, or production confirmation.return_to_rework: claimed completion lacks evidence or violates the journey contract.skip_with_reason: stage is intentionally skipped with accepted reason and visible impact.block: continuing would compromise quality, privacy, legality, production safety, or journey integrity.
Hard Rules
- Do not accept
model_convenience,implicit_assumption,planned_counted_as_ran, ortool_not_called_but_marked_done. - Do not mark planned commands as executed.
- Do not claim that a full journey is complete when only a demo slice ran.
- Do not store secrets, tokens, cookies,
.env, recovery codes, private screenshots, or raw provider payloads. - Optional external tools, crawls, uploads, paid APIs, production deploys, analytics, ads, and account mutations require explicit approval.
Local Validation Commands
bash
python3 "$UIUX_SKILL_SYSTEM_ROOT/scripts/validate_workflow_compliance.py" --fixtures
python3 "$UIUX_SKILL_SYSTEM_ROOT/scripts/validate_workflow_log.py" --fixtures
python3 "$UIUX_SKILL_SYSTEM_ROOT/scripts/validate_workflow_compliance.py" --project "$PROJECT_ROOT" --journey new-premium-site --out "$PROJECT_ROOT/reports/workflow-compliance-report.md" --log "$PROJECT_ROOT/workflow-log.jsonl"Skill package checks:
bash
python3 $CODEX_HOME/skills/.system/skill-creator/scripts/quick_validate.py $CODEX_HOME/skills/workflow-compliance-supervisor
python3 $CODEX_HOME/skills/senior-skill-architect/scripts/lint_production_skill.py $CODEX_HOME/skills/workflow-compliance-supervisorHandoff Shape
text
Declared journey:
Checked artifacts:
Decision:
Reason code:
Evidence:
User-visible message:
Next action:
Can continue:
Must rework:
Skipped with accepted reason:
Blocked:
