本文へ移動
cccskills
無料GitHub で公開

vc-audit-plans

Audit active project plan files for staleness, completion, and routing truth. Use when cleaning up plans, reconciling active work, or archiving completed artifacts.

インストール方法を見る

含まれるファイル(18)

  • SKILL.md3.0 KB
  • references/audit-plans.md4.7 KB
  • scripts/eval-coverage-known-gaps.json7.6 KB
  • scripts/fixtures/validate-backlog-notes/fail/malformed_NOTE.md137 B
  • scripts/fixtures/validate-backlog-notes/pass/d2-fix-deferred_NOTE.md268 B
  • scripts/fixtures/validate-phase-plan-completeness/fail.md246 B
  • scripts/fixtures/validate-phase-plan-completeness/pass.md307 B
  • scripts/fixtures/validate-phase-reports/fail/umbrella.md327 B
  • scripts/fixtures/validate-phase-reports/pass/phase-01_REPORT.md71 B
  • scripts/fixtures/validate-phase-reports/pass/umbrella.md327 B
  • scripts/fixtures/validate-umbrella-state/fail.md397 B
  • scripts/fixtures/validate-umbrella-state/pass.md404 B
  • scripts/validate-backlog-notes.mjs2.0 KB
  • scripts/validate-eval-coverage.mjs9.7 KB
  • scripts/validate-phase-plan-completeness.mjs1.6 KB
  • scripts/validate-phase-reports.mjs3.5 KB
  • scripts/validate-plan-inventory.mjs3.8 KB
  • scripts/validate-umbrella-state.mjs3.3 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Audit Plans

Output style: Follow process/development-protocols/communication-standards.md — answer-first, plain language, no unexplained jargon, TL;DR on long responses.

Use this skill to review active plan artifacts and reconcile them with the current codebase.

This is a maintenance and recovery skill, not an automatic post-task hook.

Optional input: a feature, folder, plan filename, or maintenance scope to prioritize.

Prefer it when:

  • UPDATE PROCESS was skipped and active-plan cleanup drift accumulated
  • the user wants a periodic active-plan cleanup pass
  • multiple active plans need reconciliation after a burst of work

Workflow

  1. Read references/audit-plans.md for the full audit process.
  2. Run the inventory validator:
    node .claude/skills/vc-audit-plans/scripts/validate-plan-inventory.mjs
    
  3. Inventory plans in process/general-plans/active/ and process/features/*/active/. Plans now live inside {slug}_{date}/ task subfolders — scan one level deep. Do NOT count _REPORT_, _REF_, or _SPEC_ files inside task folders as plans; only _PLAN_ files count. For feature-scoped audits, first run find process/features/{feature}/ -type f | sort for full artifact visibility. For full audits, run find process/features/ -type f | sort to see all feature artifacts across all subdirs (active, completed, backlog, references, reports). 3.5. Scan task folder contents (co-located REPORT/REF/SPEC files) alongside each plan. Per task-folder artefact colocation, the correct home for every artefact (plan, spec, reports, references) is INSIDE its {slug}_{date}/ task folder; flag any task artefact found in the deprecated sibling reports//references/ dirs or any ad-hoc location as mis-located, and recommend moving it into the owning task folder. Match by feature slug, date proximity (7 days), or content reference to the plan filename.
  4. Cross-check each plan against the actual codebase with file existence checks and targeted rg searches.
  5. Classify each plan as Completed, Partially Done, Obsolete, Stale, Active, or Reference.
  6. Move only clearly completed or obsolete plans to the appropriate completed/ folder. Use git mv active/{slug}_{date}/ completed/{slug}_{date}/ — move the WHOLE task folder; no completed_ prefix added.
  7. Ask before deleting anything.
  8. Re-run the inventory validator after moving or editing plan files.

Output

Return a concise summary table with classification, action taken, and any user decisions needed. Include stale artifact findings (reports/references tied to completed or obsolete plans) with recommended actions.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

AI-optimized browser automation CLI with context-efficient snapshots. Use for long autonomous sessions, self-verifying workflows, video recording, and cloud browser testing (Browserbase).

日本語の概要は準備中です。原文の説明を表示しています。

withkynam/vibecode-pro-max-kit1,1472026年6月21日 更新

Evaluate 4 execution strategies (sequential, parallel-subagents, workflow, agent-team) for a phase or fan-out task. Outputs 7-signal score table, agent count math, cost guards, and strategy recommendation.

日本語の概要は準備中です。原文の説明を表示しています。

withkynam/vibecode-pro-max-kit1,1472026年6月21日 更新

Audit project context routing, shared-skill discoverability, and Claude/Codex wiring. Use when context docs or skill surfaces move, split, or drift.

日本語の概要は準備中です。原文の説明を表示しています。

withkynam/vibecode-pro-max-kit1,1472026年6月21日 更新

Audit agent harness health: Claude/Codex agent parity, skill registry consistency, README.md sync, and protocol file wiring. Use when agents, skills, README.md, or development-protocol files move, split, or drift.

日本語の概要は準備中です。原文の説明を表示しています。

withkynam/vibecode-pro-max-kit1,1472026年6月21日 更新

Emit and validate the provisional goal block for Autopilot Mode. Owns the 9-field format and resume detection from a pasted goal block.

日本語の概要は準備中です。原文の説明を表示しています。

withkynam/vibecode-pro-max-kit1,1472026年6月21日 更新

Loop: find gaps → fix → repeat until agents find no gaps or a metric goal is hit. Shared loop primitive for PVL, EVL, and standalone quality runs.

日本語の概要は準備中です。原文の説明を表示しています。

withkynam/vibecode-pro-max-kit1,1472026年6月21日 更新

withkynam のスキルをすべて見る

このスキルの問題を報告する