본문으로 건너뛰기
cccskills
무료GitHub에 공개

vcsdd-harden

Run Phase 5 (formal hardening) for the active VCSDD feature. Invokes vcsdd-verifier to execute proof obligations, security hardening, and a purity-boundary audit before convergence.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(1)

  • SKILL.md3.7 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

What

Runs formal hardening (Phase 5). Invokes the vcsdd-verifier agent to execute language-appropriate verification tools against the proof obligations defined in Phase 1b, run security hardening checks, and audit the purity boundary. Produces verification-report.md, security-report.md, and purity-audit.md.

When

Run once the feature is already at phase 5. This happens either after adversarial review PASS (3 -> 5) or after Phase 4 explicitly routes the current sprint's findings to Phase 5 (3 -> 4 -> 5) for proof-gap / invariant-only hardening work.

How

  1. Read proof obligations from state.json.proofObligations
  2. Filter required obligations: skip Tier 0 and non-required obligations
  3. Run security hardening regardless of proof count:
    • execute Semgrep/Wycheproof/project-appropriate equivalents when relevant
    • capture raw outputs under verification/security-results/
    • summarize results in verification/security-report.md
  4. Audit purity boundaries regardless of proof count:
    • compare the implemented core/shell split against specs/verification-architecture.md
    • write findings and residual risks to verification/purity-audit.md
  5. If no required obligations: still write a lightweight verification report that documents the absence of required proofs, the security sweep, and the purity audit
  6. Invoke vcsdd-verifier agent for each obligation:
    • Write proof harness to verification/proof-harnesses/
    • Run appropriate tool (Kani, hypothesis, fast-check)
    • Capture results to verification/fuzz-results/ or verification/mutation-results/
    • Update obligation status in state.json
  7. Write verification-report.md with proof results and any graceful degradation
    • required sections: ## Proof Obligations, ## Summary
    • required obligations must end as proved; a required obligation left as skipped blocks Phase 6
  8. Write security/purity artifacts with required structure
    • verification/security-report.md must include ## Tooling and ## Summary
    • verification/purity-audit.md must include ## Declared Boundaries, ## Observed Boundaries, and ## Summary
    • verification/security-results/ must contain at least one captured output file, even if the tools were not applicable
  9. Transition to Phase 6 only when:
    • all required obligations are proved
    • verification/verification-report.md exists and was written after entering Phase 5
    • verification/security-report.md exists and was written after entering Phase 5
    • verification/purity-audit.md exists and was written after entering Phase 5
    • verification/security-results/ contains at least one captured output artifact written after entering Phase 5

Language profile resolution

  1. state.json.language (canonical; set by initFeature(..., language) at /vcsdd-init)
  2. Else .vcsdd/index.json → features.<name>.language (denormalized cache)
  3. Else treat as unspecified — verifier should infer from the repo (lockfiles, extensions) or ask the human.

Use getLanguageForFeature(featureName) from scripts/lib/vcsdd-state.js for (1)+(2).

Load tool hints from the installed plugin copy of manifests/language-profiles.json (tiers, install commands, red/green/coverage commands) for the resolved language.

Examples

/vcsdd-harden
/vcsdd-harden --tier 1    # run only Tier 1 tools (property tests/fuzzing)
/vcsdd-harden --skip-optional    # skip non-required obligations

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Use this skill when running or interpreting VCSDD adversarial reviews. Provides calibration guidance for the Adversary agent, finding severity classification, and anti-leniency enforcement patterns.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

sc30gsw/vcsdd-claude-code272026년 4월 5일 업데이트

Run Phase 3 (adversarial review) for the active VCSDD feature. Spawns a fresh vcsdd-adversary agent with zero Builder context to review implementation against spec. Produces binary PASS/FAIL verdict per dimension.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

sc30gsw/vcsdd-claude-code272026년 4월 5일 업데이트

Create a git commit for the current VCSDD phase. Auto-generates a conventional commit message from state.json including phase, sprint, gate verdict, and affected bead IDs. Tags with vcsdd/<feature>/phase-<id>.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

sc30gsw/vcsdd-claude-code272026년 4월 5일 업데이트

Run the strict-mode sprint contract review after Phase 2c. Spawns a fresh vcsdd-adversary instance to review `contracts/sprint-N.md` and writes a PASS/FAIL verdict under `reviews/contracts/sprint-N/output/`.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

sc30gsw/vcsdd-claude-code272026년 4월 5일 업데이트

Run Phase 6 (convergence check) for the active VCSDD feature. Validates four-dimensional convergence: finding diminishment, finding specificity, criteria coverage, and duplicate detection.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

sc30gsw/vcsdd-claude-code272026년 4월 5일 업데이트

Use this skill during Phase 6 convergence checks. Provides four-dimensional convergence analysis, hallucination detection, and duplicate finding identification.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

sc30gsw/vcsdd-claude-code272026년 4월 5일 업데이트

sc30gsw의 스킬 모두 보기

이 스킬의 문제 신고