Agent Skills: Quality Checks

Quality Checks

Run comprehensive quality gate checks on code changes. Orchestrates quality checks at key lifecycle points in the development workflow.

When to Use This Skill

• Before starting significant code changes (pre-edit baseline)
• After completing code changes (post-edit validation)
• Before marking a task as complete (final quality gate)
• When asked to run "quality check" or "quality gate"

Skill Contents

Sections

• When to Use This Skill
• Sub-Skills
• Lifecycle Hooks
• IDE Integration
• Running Quality Checks
• Commands
• Assets
• References
• Test Result Caching
• Related Skills

Available Resources

📚 references/ - Detailed documentation

• test result caching

📦 assets/ - Templates and resources

• claude quality hooks
• cursor quality integration

Sub-Skills

Quality checks orchestrate these sub-skills:

| Sub-Skill | Purpose | Skill Location | |-----------|---------|----------------| | testing-standards | Validates test coverage | SKILL.md | | doc-validation-rfc-37 | Validates documentation | SKILL.md | | coding-standards | Enforces code style | SKILL.md | | security-review | Checks for vulnerabilities | SKILL.md |

Lifecycle Hooks

| Hook | When | Purpose | |------|------|---------| | pre-edit | Before changes | Capture baseline metrics | | post-edit | After changes | Validate changes meet standards | | on-stop | Before completion | Final quality gate |

IDE Integration

The quality gateway integrates with AI IDEs through hooks:

| IDE | Integration Method | Reference | |-----|-------------------|-----------| | Claude Code | Native hooks in settings.json | assets/claude-quality-hooks.md | | Cursor IDE | Rules + Commands | assets/cursor-quality-integration.md |

For hook implementation patterns, see the ai-agent-hooks skill.

Running Quality Checks

Pre-Edit (Baseline)

Before making changes:

1. Record current test coverage
2. Note existing linting errors
3. Capture documentation state

Post-Edit (Validation)

After making changes:

1. Verify test coverage maintained or improved
2. Check for new linting errors
3. Validate documentation is in sync

On-Stop (Final Gate)

Before completing:

1. Run full test suite
2. Verify no regressions
3. Check all quality thresholds met

Commands

| Command | Purpose | |---------|---------| | /quality-check | Run full quality gate | | /add-tests | Generate missing tests | | /sync-docs | Update documentation |

Assets

| Asset | Description | |-------|-------------| | assets/claude-quality-hooks.md | Claude Code hook configurations | | assets/cursor-quality-integration.md | Cursor IDE integration guide |

References

| Reference | Description | |-----------|-------------| | references/test-result-caching.md | How to save and reuse test results |

Test Result Caching

Always save test output to .tmp/ instead of running tests multiple times:

# Run once and save
pnpm test 2>&1 | tee .tmp/pnpm-test-latest.txt

# Then grep from file (don't run tests again!)
grep "FAIL" .tmp/pnpm-test-latest.txt
tail -20 .tmp/pnpm-test-latest.txt

See references/test-result-caching.md for all tools.

Related Skills

| Skill | Purpose | |-------|---------| | ai-agent-hooks | Hook implementation patterns | | testing-standards | Test coverage validation | | doc-validation-rfc-37 | Documentation validation | | coding-standards | Code style enforcement | | security-review | Security vulnerability checks |