Claude Code Plugins

Community-maintained marketplace

Feedback

quality-gate-validation

@frostaura/ai.toolkit.gaia
23
0

Guide for running and validating quality gates (build, lint, test, coverage, regression). Use when validating code, running tests, or checking if a phase can proceed.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name quality-gate-validation
description Guide for running and validating quality gates (build, lint, test, coverage, regression). Use when validating code, running tests, or checking if a phase can proceed.

Quality Gate Validation

Gate Commands

Gate .NET Node.js
Build dotnet build npm run build
Lint dotnet format --verify-no-changes npm run lint
Test dotnet test npm test / npx playwright test
Coverage dotnet test --collect:"XPlat Code Coverage" npm test -- --coverage

Coverage Requirement

100% code coverage required for both frontend and backend. No exceptions.

Validation Rule

All gates are binary pass/fail - exit code 0 = pass, anything else = fail.

Retry Strategy

Attempt 1: Fix the identified issue → re-run gate
Attempt 2: Simplify the approach → re-run gate
Attempt 3: Suggest architectural refactor to retain functionality but simplify testability of the code -> get user approval -> implement the changes → re-run gate
Attempt 4: Reduce scope (remove problematic feature) → re-run gate
Still failing: Mark task as "blocked", store reason, continue with other tasks

Phase 7 Validation Checklist

After each feature, ALL must pass:

  1. Test Suite: Unit + Integration + E2E pass (exit 0)
  2. Coverage: 100% code coverage (frontend + backend)
  3. Functional Regression: Manually test features with Playwright (not spec files)
  4. Visual Regression: Playwright screenshots match baseline
  5. Performance: Response times within 5% of baseline
  6. User Journeys: All existing workflows functional

Functional Regression (Manual Testing)

Use Playwright interactively to verify features work - this is NOT writing spec files:

  • Navigate to each affected page/component
  • Test all user interactions (clicks, forms, navigation)
  • Verify data displays correctly
  • Check error states and edge cases
  • Confirm no console errors

If Validation Fails

  1. STOP development immediately
  2. Root cause analysis
  3. Fix or redesign approach
  4. Re-validate until 100% pass
  5. Store: remember("regression", "feature_x_issue", "[details]")

Store Gate Results

# On success
mcp__gaia__remember("gate", "phase_X", "passed", "ProjectWide")

# On blocked (after 3 attempts)
mcp__gaia__remember("gate", "phase_X_blocked", "[reason and what was tried]", "ProjectWide")

Gate Summary

Gate Validates Failure Action
Build Compilation Fix syntax/type errors
Lint Code style Run auto-fix or manual fix
Test Functionality Debug failing tests
Coverage 100% coverage Add missing tests
Functional Features work Fix broken functionality
Regression No breakage Investigate root cause