ccpm-code-review

ccpm-code-review

Enforces quality verification gates with four-step validation (tests pass, build succeeds, checklist complete, no blockers) before task completion, PR creation, or status updates. Auto-activates when user says "done", "complete", "finished", "ready to merge", or runs /ccpm:verify or /ccpm:done commands. Provides systematic verification workflow that prevents false completion claims and ensures production readiness. Blocks external system writes (Jira, Slack) until evidence collected. Integrates with external-system-safety for confirmation workflow. When verification fails, suggests /ccpm:verify to debug issues systematically.

6звезд
1форков
Обновлено 1/12/2026
SKILL.md
readonlyread-only
name
ccpm-code-review
description

Enforces quality verification gates with four-step validation (tests pass, build succeeds, checklist complete, no blockers) before task completion, PR creation, or status updates. Auto-activates when user says "done", "complete", "finished", "ready to merge", or runs /ccpm:verify or /ccpm:done commands. Provides systematic verification workflow that prevents false completion claims and ensures production readiness. Blocks external system writes (Jira, Slack) until evidence collected. Integrates with external-system-safety for confirmation workflow. When verification fails, suggests /ccpm:verify to debug issues systematically.

CCPM Code Review

Structured code review workflow integrated with CCPM's Linear-based project management system. Enforces "no completion claims without verification evidence" principle.

When to Use

This skill auto-activates when:

  • User says "done", "complete", "finished", "ready to merge"
  • Running /ccpm:verify command (natural workflow - recommended)
  • Running /ccpm:done command (includes pre-flight verification)
  • Running /ccpm:verify command (advanced)
  • Before updating Linear task status to "Done"
  • Before syncing Jira status
  • Before creating GitHub/BitBucket PR
  • Before sending Slack completion notifications

CCPM Verification Workflow

CCPM provides a streamlined 3-step verification process:

Step 1: Quality Checks (Automated)

Run linting, tests, and build checks to ensure technical correctness:

/ccpm:verify [issue-id]  # Auto-detects issue from git branch

What it checks:

  • ✅ Linting passes (no style errors)
  • ✅ Tests pass (all unit and integration tests)
  • ✅ Build succeeds (no compilation errors)
  • ✅ Checklist complete (100% of implementation items)

If checks fail: Command automatically suggests /ccpm:verify to debug systematically.

Step 2: Agent Code Review

After quality checks pass, agent review analyzes:

  • ✅ Code quality and best practices
  • ✅ Security vulnerabilities
  • ✅ Performance implications
  • ✅ Requirement fulfillment
  • ✅ Regression risks

Step 3: Final Confirmation

Four verification gates must pass:

  1. Tests Pass ✅ - Zero failures
  2. Build Succeeds ✅ - Exit status 0
  3. Checklist Complete ✅ - 100% checked
  4. No Blockers ✅ - No unresolved blockers in Linear

Only after all gates pass can task proceed to finalization with /ccpm:done.

Core Principles

1. Technical Correctness Over Social Comfort

Forbidden performative agreement:

  • ❌ "Great point!"
  • ❌ "You're absolutely right!"
  • ❌ "That makes total sense!"

Required rigorous verification:

  • ✅ "Let me verify that assumption"
  • ✅ "I'll test this approach first"
  • ✅ "Here's the evidence: [test output]"

2. No Implementation Before Verification

When receiving feedback:

  1. First: Verify the feedback is technically correct
  2. Then: Implement the change
  3. Never: Blindly implement without understanding

3. NO COMPLETION CLAIMS WITHOUT EVIDENCE

Required evidence before any "done" claim:

  • ✅ Tests: All passing (screenshot or CI link)
  • ✅ Build: Exit status 0 (no errors)
  • ✅ Linear checklist: 100% complete
  • ✅ No unresolved blockers in Linear comments

Integration with CCPM Commands

With /ccpm:verify (Natural Workflow - Recommended)

This skill enforces verification during the streamlined verify command:

# User runs natural verification command
/ccpm:verify

# Or with explicit issue ID
/ccpm:verify AUTH-123

Flow:

User: "I'm done, let me verify"

Claude: [ccpm-code-review activates]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔍 Smart Verify Command
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📋 Issue: AUTH-123 - Add user authentication
📊 Status: In Progress
📋 Checklist: 100% (5/5 items)

Verification Flow:
──────────────────
1. Quality Checks (linting, tests, build)
2. Final Verification (code review, security)

═══════════════════════════════════════
Step 1/2: Running Quality Checks
═══════════════════════════════════════

🔍 Running linting...
✅ All files pass linting

🧪 Running tests...
✅ All tests passed (28/28)

🏗️  Running build...
✅ Build successful

📊 Quality Check Results:
  ✅ Linting
  ✅ Tests
  ✅ Build

═══════════════════════════════════════
Step 2/2: Running Final Verification
═══════════════════════════════════════

[Code reviewer agent analyzes changes...]

✅ All requirements met
✅ Code quality standards met
✅ Security best practices followed
✅ Performance acceptable

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✅ All Verification Complete!
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

All verifications passed! Ready to finalize.

💡 What's Next?
⭐ Recommended: /ccpm:done AUTH-123

If verification fails:

❌ Quality Checks Failed

📊 Quality Check Results:
  ✅ Linting
  ❌ Tests (3 failures)
  ✅ Build

To debug and fix issues:
  /ccpm:verify AUTH-123

Then run verification again:
  /ccpm:verify AUTH-123

With /ccpm:done (Pre-Flight Verification)

This skill ensures quality before finalization:

# User attempts to finalize task
/ccpm:done

# Or with explicit issue ID
/ccpm:done AUTH-123

Pre-flight checks enforced by ccpm-code-review:

User: "/ccpm:done AUTH-123"

Claude: [ccpm-code-review activates for pre-flight]

✅ All pre-flight checks passed!

Checking completion readiness:

□ Tests passing?
  → ✅ All 52 tests passed

□ Build successful?
  → ✅ Build completed (exit 0)

□ Linear checklist complete?
  → ✅ 7/7 items checked

□ No blockers?
  → ✅ No blockers found

VERIFICATION COMPLETE ✅

📝 Creating GitHub Pull Request...
✅ Pull Request created

[Prompts for Jira/Slack updates with confirmation]

✅ Linear issue updated to Done

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🎉 Task Finalized: AUTH-123
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

If pre-flight fails:

⛔ Cannot Finalize: Checklist Incomplete

Progress: 80% (4/5 completed)

❌ Remaining Items:
  - [ ] Write integration tests

🔧 Actions Required:
1. Complete remaining checklist items
2. Update checklist: /ccpm:sync AUTH-123
3. Then run: /ccpm:done AUTH-123

Advanced: With /ccpm:verify

For manual verification step (advanced users):

Claude runs /ccpm:verify AUTH-123

[ccpm-code-review activates during verification]

Checking verification evidence:

□ Tests passing?
  → ✅ All 47 tests passed

□ Build successful?
  → ✅ Build completed with exit code 0

□ Linear checklist complete?
  → ❌ 3 of 5 items checked

□ No blockers?
  → ⚠️ 1 blocker found: "Redis connection failing in staging"

VERIFICATION FAILED:
- Linear checklist incomplete (3/5)
- Blocker unresolved: Redis connection issue

Cannot proceed with completion until:
1. Complete remaining checklist items
2. Resolve Redis blocker or document as out-of-scope

Would you like to:
A. Update checklist now
B. Investigate Redis blocker
C. Update task scope to exclude Redis

Receiving Feedback Protocol

When PR Reviewer Comments

Scenario: Reviewer says "Add error handling to login function"

❌ Wrong approach (performative):

"Great point! I'll add error handling right away."
[Immediately adds try-catch without understanding]

✅ Correct approach (verification-first):

Claude: [ccpm-code-review activates]

Let me verify this feedback:

1. Checking current error handling...
   ```typescript
   // login.ts
   async function login(email, password) {
     const user = await db.findUser(email);  // ← No error handling
     return generateToken(user);             // ← No error handling
   }
  1. Potential issues:

    • Database query could fail (network, timeout)
    • user might be null (not found)
    • generateToken could throw (invalid data)
  2. Verification: Feedback is VALID

    • Missing error handling confirmed
    • Could cause uncaught exceptions
  3. Implementation approach:

    async function login(email, password) {
      try {
        const user = await db.findUser(email);
        if (!user) {
          throw new UnauthorizedError('Invalid credentials');
        }
        return generateToken(user);
      } catch (error) {
        if (error instanceof DatabaseError) {
          logger.error('Database error during login', error);
          throw new ServiceUnavailableError();
        }
        throw error;
      }
    }
    
  4. Testing the fix:

    • Add test: login with network error → returns 503
    • Add test: login with invalid user → returns 401
    • Verify existing tests still pass
  5. Update Linear:

    • Add comment: "PR feedback: Added error handling"
    • Link to new commit

Implementing fix now...


### When Feedback Seems Wrong

**Scenario**: Reviewer says "Use var instead of const"

**❌ Wrong approach**:

"Okay, I'll change it!"
[Blindly makes change]


**✅ Correct approach**:

Claude: [ccpm-code-review activates]

Let me verify this feedback:

  1. Current code uses const (ES6+)

  2. Feedback suggests var (ES5)

  3. Technical analysis:

    • const: Block-scoped, immutable binding, prevents reassignment bugs
    • var: Function-scoped, allows reassignment, hoisting issues
  4. Project standards check:

    • .eslintrc: Enforces const/let, disallows var
    • Existing codebase: 100% uses const/let
  5. Verification: Feedback appears INCORRECT

    • Contradicts project standards
    • Would cause linting failures
    • Downgrades code quality
  6. Response to reviewer:
    "I checked our ESLint config and existing codebase. We enforce const/let
    and disallow var for block-scoping safety. Changing to var would:

    • Fail CI linting
    • Contradict project standards
    • Introduce potential hoisting bugs

    Did you mean something else, or is there a specific reason for var here?"

Waiting for reviewer clarification before making changes...


## Verification Gates

### Gate 1: Tests Must Pass

**Requirement**: Zero test failures

✅ PASS: 52 tests, 0 failures
❌ FAIL: 48 tests passed, 4 failures

If failures exist:

  1. Run /ccpm:verify to debug
  2. Update Linear with findings
  3. Re-run verification after fixes
  4. Only proceed when 0 failures

### Gate 2: Build Must Succeed

**Requirement**: Exit status 0

✅ PASS: Build completed successfully (exit 0)
❌ FAIL: Build failed with 3 TypeScript errors (exit 1)

If build fails:

  1. Review error messages
  2. Fix compilation errors
  3. Re-run build
  4. Only proceed when exit 0

### Gate 3: Linear Checklist Must Be Complete

**Requirement**: 100% of checklist items checked

✅ PASS: 7/7 checklist items complete
❌ FAIL: 5/7 checklist items complete

Incomplete items:

  • [ ] Add integration tests
  • [ ] Update API documentation

Action required:

  1. Complete remaining items, OR
  2. Update task scope (with /ccpm:plan)
  3. Mark items as out-of-scope explicitly
  4. Only proceed when justified

### Gate 4: No Unresolved Blockers

**Requirement**: Zero blocker comments in Linear

✅ PASS: No blockers found
❌ FAIL: 2 blockers in Linear comments

Blockers:

  1. "Database migration failing in staging"
  2. "Awaiting security team approval"

Action required:

  1. Resolve each blocker, OR
  2. Document why blocker doesn't apply
  3. Get explicit approval to proceed despite blocker
  4. Only proceed when all blockers addressed

## Integration with CCPM Safety Rules

### Works with `external-system-safety` Skill

This skill focuses on **verification**, `external-system-safety` focuses on **confirmation**.

**Division of responsibility:**

| Concern | Skill | Check |
|---------|-------|-------|
| Are tests passing? | ccpm-code-review | ✓ |
| Is build successful? | ccpm-code-review | ✓ |
| Is checklist complete? | ccpm-code-review | ✓ |
| User confirms Jira update? | external-system-safety | ✓ |
| User confirms Slack post? | external-system-safety | ✓ |

**Together they ensure**:
1. Quality standards met (ccpm-code-review)
2. User approves external writes (external-system-safety)

### References SAFETY_RULES.md

All external writes follow `../commands/SAFETY_RULES.md`:

**⛔ Require confirmation**:
- Jira status updates
- Confluence page edits
- BitBucket PR creation
- Slack notifications

**✅ Always allowed**:
- Linear status updates (internal tracking)
- Local file modifications
- Read operations

## Examples

### Example 1: Successful Completion Flow (Natural Workflow)

User: "Done with AUTH-123, ready to ship!"

Claude: [ccpm-code-review activates]

Let me run verification first:

Running /ccpm:verify AUTH-123...

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔍 Smart Verify Command
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📋 Checklist: 100% (7/7 items)

═══════════════════════════════════════
Step 1/2: Running Quality Checks
═══════════════════════════════════════

✅ Tests: 52/52 passing
✅ Build: Successful (exit 0)
✅ Linting: No errors

═══════════════════════════════════════
Step 2/2: Running Final Verification
═══════════════════════════════════════

✅ Code review: Passed
✅ Security audit: Passed
✅ Requirements met: Passed

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✅ All Verification Complete!
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

💡 What's Next?
⭐ Recommended: /ccpm:done AUTH-123

Ready to finalize! Shall I run /ccpm:done now?


### Example 2: Failed Verification (Natural Workflow)

User: "I'm done, let's merge"

Claude: [ccpm-code-review activates]

Running /ccpm:verify...

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔍 Smart Verify Command
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

═══════════════════════════════════════
Step 1/2: Running Quality Checks
═══════════════════════════════════════

🧪 Running tests...
❌ 3 tests failing:

  • test/auth/login.test.ts: "should handle invalid credentials"
  • test/auth/jwt.test.ts: "should refresh expired tokens"
  • test/api/protected.test.ts: "should reject missing tokens"

🏗️ Running build...
❌ TypeScript error:
src/auth/jwt.ts:45:12 - error TS2345

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
❌ Verification Failed
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

To debug and fix issues:
/ccpm:verify AUTH-123

Then run verification again:
/ccpm:verify AUTH-123

Would you like me to start debugging these issues?


### Example 3: Incomplete Checklist (Pre-Flight with /ccpm:done)

User: "/ccpm:done AUTH-456"

Claude: [ccpm-code-review activates for pre-flight]

✅ All git pre-flight checks passed!

Checking completion readiness:

✅ Tests passing (41/41)
✅ Build successful
❌ Linear checklist incomplete (4/6):
✓ Implement JWT generation
✓ Add login endpoint
✓ Add logout endpoint
✓ Write unit tests
✗ Add integration tests
✗ Update API documentation

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⛔ Cannot Finalize: Checklist Incomplete
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Progress: 67% (4/6 completed)

🔧 Actions Required:

  1. Complete remaining checklist items
  2. Update checklist: /ccpm:sync AUTH-456
  3. Then run: /ccpm:done AUTH-456

Or update scope with /ccpm:plan AUTH-456 "mark items as Phase 2"

Which would you like to do?


### Example 4: Complete Natural Workflow (End-to-End)

User: "I finished the authentication feature"

Claude: [ccpm-code-review guides through workflow]

Great! Let's verify your work before finalizing.

Step 1: Verification
────────────────────
/ccpm:verify

✅ All quality checks passed
✅ Code review complete
✅ Security validated

Step 2: Finalization
────────────────────
/ccpm:done

✅ Git pre-flight checks passed
✅ Checklist complete (5/5)
✅ No blockers
📝 GitHub PR created
✅ Linear updated to Done

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🎉 Task Finalized: AUTH-456
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

All done! Task completed successfully. 🎊


## Verification Checklist Template

Before claiming "done":

- [ ] Tests executed
  - [ ] All tests passing (0 failures)
  - [ ] Coverage meets requirements
  - [ ] New tests added for new code

- [ ] Build verified
  - [ ] Build successful (exit 0)
  - [ ] No compilation errors
  - [ ] No linting errors

- [ ] Linear task complete
  - [ ] All checklist items checked
  - [ ] No unresolved blockers
  - [ ] Work summary added

- [ ] Code quality
  - [ ] Code reviewed (by human or code-reviewer agent)
  - [ ] Security checked (if applicable)
  - [ ] Performance acceptable

- [ ] Documentation
  - [ ] Code comments added where needed
  - [ ] API docs updated (if API changed)
  - [ ] README updated (if user-facing)

Only after ALL checked:
- [ ] Ready for /ccpm:done

## Integration with Other CCPM Skills

**Works alongside**:

- **external-system-safety**: Enforces confirmation for external writes (Jira/Slack)
- **pm-workflow-guide**: Suggests `/ccpm:verify` and `/ccpm:done` at right time
- **ccpm-debugging**: Invoked via `/ccpm:verify` when checks fail
- **sequential-thinking**: For complex verification scenarios

**Example combined activation**:

User: "Ready to merge AUTH-123"

ccpm-code-review → Suggests /ccpm:verify first

/ccpm:verify → Runs quality checks + agent review

[If gates pass]

Suggests /ccpm:done

/ccpm:done → Pre-flight checks + PR creation

external-system-safety → Confirms Jira/Slack writes

[If user confirms]

Complete! ✅


## Natural Workflow Commands

CCPM provides streamlined commands for the complete verification and finalization workflow:

| Command | Purpose | Auto-detects Issue |
|---------|---------|-------------------|
| `/ccpm:verify` | Quality checks + agent review | ✅ From git branch |
| `/ccpm:done` | Pre-flight + PR + finalize | ✅ From git branch |
| `/ccpm:verify` | Debug failed checks | ❌ Explicit ID required |

**Recommended workflow:**
```bash
# 1. Complete implementation
/ccpm:work

# 2. Commit changes
/ccpm:commit

# 3. Verify quality (this skill activates)
/ccpm:verify

# 4. Finalize task (this skill activates for pre-flight)
/ccpm:done

Summary

This skill ensures:

  • ✅ No false completion claims
  • ✅ Evidence required before "done"
  • ✅ Quality gates enforced (4-step validation)
  • ✅ Technical rigor over social comfort
  • ✅ Integration with CCPM natural workflow
  • ✅ Systematic debugging when failures occur

Philosophy: Verification before completion, evidence over claims, quality over speed.

Key Features:

  • Auto-activates on completion attempts
  • Enforces 4 verification gates
  • Integrates with /ccpm:verify and /ccpm:done
  • Suggests /ccpm:verify for failures
  • Works with external-system-safety for confirmations

Source: Adapted from claudekit-skills/code-review
License: MIT
CCPM Integration: /ccpm:verify, /ccpm:done, /ccpm:verify, quality-gate hook

You Might Also Like

Related Skills

fix

fix

243Kdev-testing

Use when you have lint errors, formatting issues, or before committing code to ensure it passes CI.

peekaboo

peekaboo

179Kdev-testing

Capture and automate macOS UI with the Peekaboo CLI.

frontend-testing

frontend-testing

128Kdev-testing

Generate Vitest + React Testing Library tests for Dify frontend components, hooks, and utilities. Triggers on testing, spec files, coverage, Vitest, RTL, unit tests, integration tests, or write/review test requests.

frontend-code-review

frontend-code-review

127Kdev-testing

Trigger when the user requests a review of frontend files (e.g., `.tsx`, `.ts`, `.js`). Support both pending-change reviews and focused file reviews while applying the checklist rules.

code-reviewer

code-reviewer

92Kdev-testing

Use this skill to review code. It supports both local changes (staged or working tree) and remote Pull Requests (by ID or URL). It focuses on correctness, maintainability, and adherence to project standards.

google-gemini avatargoogle-gemini
Получить
session-logs

session-logs

90Kdev-testing

Search and analyze your own session logs (older/parent conversations) using jq.