Skip to main content
ClaudeWave
Skill904 repo starsupdated yesterday

vc-generate-closeout

# vc-generate-closeout This Claude Code skill generates a structured post-execution closeout packet for completed plans or phases, including archive-readiness classification, drift signal scoring, commit checkpoint recommendations, and next-state guidance. Use it at the end of non-trivial execute completions, before archiving plans, at fast-mode-agent session end after implementation verification, or between phases in multi-phase programs before entering update process mode.

Install in Claude Code
Copy
git clone --depth 1 https://github.com/withkynam/vibecode-pro-max-kit /tmp/vc-generate-closeout && cp -r /tmp/vc-generate-closeout/.claude/skills/vc-generate-closeout ~/.claude/skills/vc-generate-closeout
Then start a new Claude Code session; the skill loads automatically.

SKILL.md

# vc-generate-closeout

> Output style: closeout packet leads with the verdict/recommendation, tables for drift/readiness, one-line TL;DR — `process/development-protocols/communication-standards.md`.

Generate the post-EXECUTE closeout packet for a completed plan or phase. Produces a structured summary with archive-readiness classification, drift signal scoring, commit checkpoint recommendation, and the single best next valid state.

## When To Invoke

Invoke this skill at the end of:

- Any non-trivial EXECUTE completion (the execute-agent's own closeout block)
- ENTER UPDATE PROCESS MODE flows before archiving a plan
- fast-mode-agent session end, after implementation and verification are complete
- Phase program closeout between phases (after validate + regression checkpoint, before inter-phase UPDATE PROCESS)

Do not invoke for trivial single-file fixes where the plan file is not involved.

## Mode Selection

Choose one mode before generating the closeout packet.

### Simple Mode (default)

Use when:
- The execute session just ended and all context is fresh in the conversation.
- The plan was a single-phase, straightforward implementation.
- No context compaction occurred during the session.

Behavior: write the 8-item closeout packet directly from the plan file and conversation context. No additional git or file scanning is required.

### Deep Mode

Trigger conditions — use Deep Mode if **any one** of the following applies:

- The session was resumed after a context compaction (conversation context may be incomplete).
- The execute session was long — many sub-steps, multiple checklist sections, or multiple commits.
- The caller explicitly requests deep mode.
- The phase is part of a multi-phase program and accurate phase-completion status is required for the umbrella plan.

Behavior: gather evidence before writing the closeout packet. Run these steps in order:

1. `git diff HEAD~1 --name-only` — confirm which files were actually changed. Cross-reference against the plan's blast radius.
2. `git log --oneline -5` — confirm commit messages match what was planned. Flag any commits that touch files outside the plan's scope.
3. Read the phase report if one was already written — confirm its claims match actual execution output.
4. Read the validate-contract test gates — confirm each gate's final status: green, known-gap, or skipped. Do not rely on memory for gate outcomes.
5. Read the plan Implementation Checklist — for every checked item, confirm a matching file appears in the git diff. Flag any checked item with no corresponding change.

After gathering evidence, write the closeout packet with explicit source citations for every material claim.

**Quality difference between modes:**

- Simple: "Section 2 completed — billing router updated" (from memory)
- Deep: "Section 2 completed — `git diff HEAD~1` confirms `packages/api/src/router/billing.ts` modified (+47/-12 lines); test gate `pnpm test:billing` confirmed green in execute log"

When in doubt, prefer Deep Mode. A false-confident Simple closeout is worse than a slightly slower Deep one.

## Closeout Packet Schema

Every closeout packet must include these 9 items. Present them in order.

1. **Selected plan path**
   - The exact file path of the plan being closed out (e.g. `process/features/foo/active/foo-phase-01_PLAN_03-06-26.md`). Never leave this implicit.

2. **Closeout classification** (one of three states — see §Closeout Classification States)

3. **What was finished**
   - A concrete list of what was actually implemented or changed. Not a restatement of the plan checklist — what was done in practice.

4. **What was verified vs still unverified**
   - What tests or evidence exist that confirm the work.
   - What still requires manual verification, user confirmation, or future test coverage.

4b. **Validate-contract compliance**
   - Was VALIDATE run for this plan?
   - Is a `## Validate Contract` section present in the plan file?
   - If VALIDATE was skipped, state the documented skip reason.
   - A plan cannot be classified `Ready for UPDATE PROCESS archival` without a present validate-contract or a documented skip reason.

5. **Cleanup done vs still needed**
   - What context docs, reports, or process artifacts were already updated.
   - What remains: open TODOs, uncommitted changes, missing reports, stale references, or plan debt.

6. **Single best next valid state** (one of the allowed states from §Move-On Semantics)
   - Name the exact next action or plan path. Never end with a generic "move to next task."

7. **Commit-checkpoint recommendation** (see §Commit Checkpoint Classification)
   - Whether to invoke `vc-git-manager` before UPDATE PROCESS, or whether the remaining changes are process-only and the commit belongs after UPDATE PROCESS.

8. **Regression status** (phase programs only)
   - Which previously verified surfaces were checked for regression against this phase's blast radius.
   - Whether all passed, or whether fixes were applied before re-verification.
   - If regression checking was skipped (e.g., first phase with no prior verified surfaces), state why explicitly.

9. **SPEC achievement**
   - For each acceptance criterion in the locked `*_SPEC_*.md`, score **met** or **unmet**.
   - Each unmet criterion → a backlog NOTE (the SPEC is frozen; gaps go to backlog only).
   - If there is no SPEC for this plan (e.g. trivial or phase-program inner loop governed by the umbrella SPEC), state that explicitly.

## Closeout Classification States

Exactly three states are allowed. Choose one and state it verbatim.

- **Ready for UPDATE PROCESS archival**
  - The selected plan path still matches the implemented work.
  - Required verification evidence exists.
  - No material deviations remain unresolved.
  - The user has confirmed or approved cleanup.
  - validate-contract is present in the plan file, or VALIDATE was explicitly skipped with a documented reason.

- **Keep in active/testing**
  - Implementation is substant
code-reviewerSubagent

Comprehensive code review with scout-based edge case detection. Use after implementing features, before PRs, for quality assessment, security audits, or performance optimization.

code-simplifierSubagent

Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality. Focuses on recently modified code unless instructed otherwise.

debuggerSubagent

Use this agent when you need to investigate issues, analyze system behavior, diagnose performance problems, examine database structures, collect and analyze logs from servers or CI/CD pipelines, run tests for debugging purposes, or optimize system performance. This includes troubleshooting errors, identifying bottlenecks, analyzing failed deployments, investigating test failures, and creating diagnostic reports. Examples:\n\n<example>\nContext: The user needs to investigate why an API endpoint is returning 500 errors.\nuser: "The /api/users endpoint is throwing 500 errors"\nassistant: "I''ll use the debugger agent to investigate this issue"\n<commentary>\nSince this involves investigating an issue, use the Task tool to launch the debugger agent.\n</commentary>\n</example>\n\n<example>\nContext: The user wants to analyze why the CI/CD pipeline is failing.\nuser: "The GitHub Actions workflow keeps failing on the test step"\nassistant: "Let me use the debugger agent to analyze the CI/CD pipeline logs and identify the issue"\n<commentary>\nThis requires analyzing CI/CD logs and test failures, so use the debugger agent.\n</commentary>\n</example>\n\n<example>\nContext: The user notices performance degradation in the application.\nuser: "The application response times have increased by 300% since yesterday"\nassistant: "I''ll launch the debugger agent to analyze system behavior and identify performance bottlenecks"\n<commentary>\nPerformance analysis and bottleneck identification requires the debugger agent.\n</commentary>\n</example>

execute-agentSubagent

EXECUTE MODE - Implementing EXACTLY what was planned. Full tool access. Can only be invoked after explicit user confirmation. Use after plan is approved.

fast-mode-agentSubagent

FAST MODE - Execute compressed RIPER-5 workflow (RESEARCH + INNOVATE + PLAN) in one session, then pause for EXECUTE confirmation. Use when you want quick end-to-end solution.

git-managerSubagent

Stage, commit, and push code changes with conventional commits. Use when user says "commit", "push", or finishes a feature/fix.

innovate-agentSubagent

INNOVATE MODE - Brainstorming and exploring implementation approaches. Discusses possibilities without making decisions. Use after research is complete.

plan-agentSubagent

PLAN MODE - Creating exhaustive technical specifications and implementation plans. Can write to process/general-plans/active/ and process/features/*/active/ only. Use after approach is decided.