Quality Gate

Purpose

Execute comprehensive quality validation before code can proceed to PR creation, ensuring all quality standards are met through automated checks and minimum thresholds.

When to Use

• Conductor workflow Phase 3 (Quality Assurance)
• Before creating any pull request
• After refactoring changes
• As part of CI/CD pipeline
• Before merging to development branch

Quality Standards

All checks must pass:

• ✅ All tests passing (unit, integration, E2E)
• ✅ Audit score ≥ 8.0/10
• ✅ Production build successful
• ✅ Test coverage ≥ 80% (configurable)
• ✅ No TypeScript errors
• ✅ No linting errors

Instructions

Step 1: Run Comprehensive Tests

Use the run-comprehensive-tests skill:

npm run test

Validation:

• Exit code 0 (all tests passed)
• No failing test cases
• Coverage data available

If tests fail:

• BLOCK - Do not proceed
• Delegate to debugger agent for investigation
• Fix failures and re-run quality gate

Step 2: Execute Code Quality Audit

Use the audit-code skill or delegate to audit agent:

Audit all changed files from this implementation:
- Architecture patterns
- Code complexity
- SOLID principles
- DRY violations
- Type safety
- Error handling

Validation:

• Audit score ≥ 8.0/10
• No CRITICAL findings
• No HIGH priority blocking issues

If audit score < 8.0:

• BLOCK - Do not proceed
• Delegate to refactor agent
• Address critical/high findings
• Re-run audit
• Must achieve ≥ 8.0 before continuing

Step 3: Validate Production Build

npm run build

Validation:

• Build completes successfully
• No compilation errors
• No build warnings (or acceptable count)
• All packages build successfully

If build fails:

• BLOCK - Do not proceed
• Investigate build errors
• Fix compilation issues
• Re-run quality gate

Step 4: Check Test Coverage

Parse coverage report from test results:

# Coverage should be included in test output
npm run test -- --coverage

Validation:

• Overall coverage ≥ 80%
• Statement coverage ≥ 80%
• Branch coverage ≥ 75%
• Function coverage ≥ 80%

If coverage below threshold:

• WARNING (not blocking for MVP)
• Log coverage gap
• Create GitHub issue for coverage improvement
• Proceed with quality gate

Step 5: TypeScript Type Checking

npm run type-check
# Or: npx tsc --noEmit

Validation:

• No TypeScript errors
• All types resolve correctly
• No implicit any types (if strict)

If type errors:

• BLOCK - Do not proceed
• Fix type errors
• Re-run quality gate

Step 6: Linting Validation

npm run lint
# Or: npx eslint .

Validation:

• No linting errors
• Warnings acceptable (log count)
• Code style consistent

If linting errors:

• Try auto-fix: npm run lint -- --fix
• Manual fix if auto-fix doesn't resolve
• Re-run quality gate

Result Aggregation

Return structured results:

{
  "qualityGate": "pass" | "fail",
  "timestamp": "2025-10-21T...",
  "checks": {
    "tests": {
      "status": "pass",
      "total": 45,
      "passed": 45,
      "failed": 0,
      "coverage": 87.5
    },
    "audit": {
      "status": "pass",
      "score": 8.5,
      "critical": 0,
      "high": 0,
      "medium": 2,
      "low": 3
    },
    "build": {
      "status": "pass",
      "duration": "12.3s",
      "warnings": 0
    },
    "typeCheck": {
      "status": "pass",
      "errors": 0
    },
    "lint": {
      "status": "pass",
      "errors": 0,
      "warnings": 2
    }
  },
  "blockers": [],
  "warnings": [
    "Coverage below target: 87.5% (target: 90%)",
    "2 lint warnings present"
  ]
}

Failure Handling

Retry Logic

If quality gate fails:

1. Identify failing check
2. Route to appropriate fix:
   • Tests fail → debugger agent
   • Audit < 8.0 → refactor agent
   • Build fails → investigate build errors
   • Type errors → fix type issues
   • Lint errors → auto-fix or manual
3. Re-run quality gate after fixes
4. Maximum 3 retries before escalating to human

Escalation Criteria

Escalate to human if:

• Quality gate fails after 3 retry attempts
• Audit score consistently < 7.0
• Build failures persist after investigation
• Tests fail with unclear error messages

Integration with Conductor

Used in Conductor Phase 3:

**Phase 3: Quality Assurance**

Use the `quality-gate` skill to validate all quality standards:

1. Execute quality gate
2. If pass: Proceed to Phase 4 (PR Creation)
3. If fail:
   - Identify failing check
   - Route to appropriate agent for fixes
   - Re-run quality gate
   - Repeat until pass or escalate

Quality gate must pass before creating PR.

Performance Optimization

Run checks in optimal order:

1. TypeScript check (fastest, catches syntax errors early)
2. Lint (fast, catches code style issues)
3. Tests (slower, but comprehensive)
4. Audit (slowest, deep analysis)
5. Build (final validation)

This order allows fast-failing for quick feedback.

Configuration

Quality gate thresholds can be adjusted:

// .claude/quality-gate-config.json
{
  "minimumAuditScore": 8.0,
  "minimumCoverage": 80,
  "blockOnCoverage": false,
  "blockOnLintWarnings": false,
  "maxRetries": 3
}

Related Skills

• run-comprehensive-tests - Test execution
• audit-code - Code quality audit
• validate-coverage - Coverage analysis
• check-lint - Linting validation
• validate-build - Build validation

Examples

See examples.md for:

• Complete quality gate execution flow
• Failure handling scenarios
• Integration with Conductor workflow
• Custom threshold configuration