GitHunt

QA Skills Project

A comprehensive skill-based system for Senior QA Engineers covering the entire software testing lifecycle.

Overview

This project implements an orchestrator-driven approach where a master coordinator delegates work to specialized QA skills. Every task goes through the orchestrator, which identifies required skills, searches for missing ones, and coordinates execution.

Installation

Install All Skills

# Install all 11 skills (requires --full-depth to discover nested skills)
npx skills add javalenciacai/QASkills --all --full-depth

Install Specific Skills

# Install individual skills (use --full-depth to access nested skills)
npx skills add javalenciacai/QASkills --skill test-design-istqb --full-depth
npx skills add javalenciacai/QASkills --skill test-execution-manager --full-depth
npx skills add javalenciacai/QASkills --skill defect-lifecycle-manager --full-depth

List Available Skills

# See all available skills before installing
npx skills add javalenciacai/QASkills --list --full-depth

Install for Specific Agents

# Install for specific AI agents
npx skills add javalenciacai/QASkills --agent claude-code --all --full-depth
npx skills add javalenciacai/QASkills --agent cursor --all --full-depth

Browse & Discover

Architecture

User Request
    ↓
senior-qa-engineer (Main Coordinator)
    ↓
multi-agent-orchestration (Master Orchestrator)
    ↓
Specialized Skills (by QA Phase)
    ↓
Results Aggregation & Delivery

Core Skills

Main Entry Point

  • senior-qa-engineer (SKILL.md): Master coordinator that defines the QA role and workflow

Master Orchestrator

  • multi-agent-orchestration: Coordinates all task execution, manages skill discovery and workflow

QA Lifecycle Skills

Phase 1: Requirements & Planning

  • user-story-verifier: Validates user stories for quality, completeness, INVEST criteria, and ISTQB applicability

Phase 2: Test Design

  • test-design-istqb: Applies ISTQB techniques (equivalence partitioning, boundary values, decision tables, state transitions) to create test cases

Phase 3: Test Execution

  • test-execution-manager: Manages test execution, defect logging, exploratory testing, and status reporting

Phase 4: Defect Management

  • defect-lifecycle-manager: Tracks defects through their lifecycle, analyzes metrics, performs root cause analysis

Phase 5: CI/CD & Automation

  • cicd-testing-integration: Integrates tests into CI/CD pipelines, enables continuous testing
  • browser-use: Automates browser interactions for web testing

Phase 6: Continuous Improvement

  • qa-process-improvement: Drives process optimization through metrics analysis and best practices

Utility Skills

  • skill-creator: Creates new specialized skills when needed
  • find-skills: Searches the skills ecosystem for capabilities

How It Works

1. Request Received

User makes a QA-related request (e.g., "Review this user story" or "Create automated tests")

2. Orchestrator Activated

The multi-agent-orchestration skill is ALWAYS activated first. It:

  • Analyzes the request
  • Identifies which QA phase(s) are involved
  • Determines required skills
  • Plans execution pattern (sequential/parallel/hierarchical)

3. Skill Discovery

For each required skill:

  • Exists locally? → Queue for execution
  • Not found? → Use find-skills to search ecosystem
    • Found? → Ask user to install
    • Not found? → Use skill-creator to propose new skill

4. Execution

Orchestrator runs skills in planned order:

  • Sequential: A → B → C (when output of A feeds B)
  • Parallel: A + B + C (when independent tasks)
  • Hierarchical: Orchestrator → Sub-tasks → Aggregation

5. Aggregation & Delivery

Orchestrator combines results and delivers comprehensive response to user

Mandatory Workflow

┌─────────────────────────────────────┐
│ 1. User Request                     │
└─────────────────────────────────────┘
                ↓
┌─────────────────────────────────────┐
│ 2. Activate multi-agent-orchestration│
└─────────────────────────────────────┘
                ↓
┌─────────────────────────────────────┐
│ 3. Analyze & Plan                   │
│    - What phase?                    │
│    - Which skills needed?           │
│    - Execution pattern?             │
└─────────────────────────────────────┘
                ↓
┌─────────────────────────────────────┐
│ 4. For each skill needed:           │
│    - Check inventory                │
│    - Search if missing              │
│    - Create if unavailable          │
└─────────────────────────────────────┘
                ↓
┌─────────────────────────────────────┐
│ 5. Execute workflow                 │
└─────────────────────────────────────┘
                ↓
┌─────────────────────────────────────┐
│ 6. Aggregate results                │
└─────────────────────────────────────┘
                ↓
┌─────────────────────────────────────┐
│ 7. Deliver to user                  │
└─────────────────────────────────────┘

Critical Rules

  1. ALWAYS activate orchestrator first - Never skip this step
  2. EVERY task uses a skill - No direct execution without skills
  3. Search before creating - Use find-skills to check ecosystem
  4. Ask before installing - Get user approval for new skills
  5. Let orchestrator decide patterns - It optimizes execution
  6. Aggregate through orchestrator - Ensures consistency

Example Scenarios

Simple Request

User: "Review this user story"
→ Orchestrator activates user-story-verifier
→ Results returned

Missing Skill

User: "Create API tests with Postman"
→ Orchestrator checks inventory: No Postman skill
→ find-skills searches ecosystem
→ Presents options to user
→ User selects, skill installed
→ Orchestrator executes with new skill

Complex Multi-Phase

User: "Review requirements and create E2E tests"
→ Orchestrator plans sequential workflow:
   Phase 1: user-story-verifier (analyze)
   Phase 2: (search for automation skill)
   Phase 3: (execute automation)
→ Aggregates results from all phases
→ Delivers comprehensive report

Benefits of This Architecture

  • Modularity: Each skill is self-contained and focused
  • Scalability: Easy to add new skills for new capabilities
  • Maintainability: Small, focused skills vs one giant file
  • Discoverability: Skills can be found and shared
  • Reusability: Skills can be used across projects
  • Consistency: Orchestrator ensures uniform execution
  • Flexibility: System adapts to missing capabilities

Skills Directory Structure

.agents/skills/
├── browser-use/
├── cicd-testing-integration/
├── defect-lifecycle-manager/
├── find-skills/
├── multi-agent-orchestration/
├── qa-process-improvement/
├── skill-creator/
├── test-design-istqb/
├── test-execution-manager/
└── user-story-verifier/

Getting Started

Using the Skills

  1. Install the skills (see Installation above)
  2. Review the main SKILL.md to understand the Senior QA Engineer role
  3. Explore individual skills in .agents/skills/ directory
  4. When making requests, let the orchestrator guide the workflow

For Contributors

  1. Fork this repository
  2. Use the skill-creator skill to generate new skill structure
  3. Place new skills in .agents/skills/[skill-name]/
  4. Follow the SKILL.md template
  5. Update skills.json with skill metadata
  6. Submit a pull request

Sharing & Publishing

Share Your Skills

Once you've created or modified skills, you can share them with the community:

  1. Push to GitHub

    git add .
    git commit -m "feat: Add new skill or update existing skills"
    git push origin main
  2. Submit to skills.sh

    • Your public GitHub repository is automatically discoverable
    • Users can install via: npx skills add yourusername/YourRepo
    • Visit skills.sh to see your skills listed
  3. Create a Release (Recommended)

    • Tag your version: git tag -a v1.0.0 -m "Release v1.0.0"
    • Push tags: git push --tags
    • Create a GitHub Release with changelog

Quality Standards

All tasks ensure:

  • ✓ Orchestrator-driven execution
  • ✓ Skill-backed operations
  • ✓ Requirements are testable
  • ✓ Test cases have clear steps and expected results
  • ✓ Defects include reproduction steps
  • ✓ Automation follows best practices
  • ✓ Coverage includes positive, negative, and edge cases
  • ✓ Traceability from requirements to tests to defects
  • ✓ Metrics drive continuous improvement

Contributing

To add a new skill:

  1. Use the skill-creator skill to generate structure
  2. Place in .agents/skills/[skill-name]/
  3. Follow SKILL.md template
  4. Update skills.json with skill metadata
  5. Update this README
  6. Test through orchestrator
  7. Submit a pull request

License

This project is licensed under the MIT License - see the LICENSE file for details.