MCP Ahrefs

--- description: Intelligently test an issue by automatically detecting whether to use MCP, API, or UI testing, and choosing between direct or subprocess execution usage: /project:test-issue [OPTIONAL-HINTS] example: /project:test-issue --- I'll analyze the current issue and automatically determine the best testing approach. ## Step 1: Analyze Context First, let me gather information to determine the testing approach: ### Check Current Branch !git branch --show-current ### Analyze Recent Changes !git diff --name-only HEAD~1..HEAD ### Look for Test Patterns !find . -name "*.test.*" -o -name "*.spec.*" -o -name "*_test.*" | head -20 ### Check Package Dependencies !test -f package.json && grep -E "(puppeteer|playwright|selenium|jest|mocha|cypress)" package.json || echo "No package.json found" !test -f requirements.txt && grep -E "(selenium|playwright|pytest)" requirements.txt || echo "No requirements.txt found" ## Step 2: Detection Logic Based on the analysis, I'll determine the testing approach: ### Indicators for UI Testing (Puppeteer): - Changes to files in: src/components/, src/pages/, src/views/, app/, frontend/ - File extensions: .jsx, .tsx, .vue, .svelte - Presence of: package.json with React/Vue/Angular/Svelte - JIRA issue mentions: UI, interface, button, form, page, screen, responsive - Existing tests using: Puppeteer, Playwright, Cypress, Selenium ### Indicators for Backend API Testing (PyTest): - Changes to files in: api/, backend/, routers/, endpoints/, controllers/ - File extensions: .py with FastAPI/Flask/Django decorators (@app.route, @router, @api_view) - Presence of: FastAPI, Flask, Django REST in requirements.txt - JIRA issue mentions: API, endpoint, REST, HTTP, request, response, validation - Existing tests using: pytest, TestClient, httpx - Files containing: test_*.py or *_test.py in tests/ directory ### Indicators for MCP Testing: - Changes to files in: mcp/, tools/ - File extensions: .py with @mcp_server decorators - Presence of: MCP server configuration - JIRA issue mentions: MCP tool, integration with external services - Current project is identified as MCP server ### Indicators for Hybrid Testing: - Full-stack changes (both frontend and backend) - JIRA issue mentions end-to-end functionality - Changes span multiple layers ## Step 3: Determine Testing Approach (Direct vs Subprocess) After detecting the test type, I'll determine whether to use direct testing or subprocess testing: ### Subprocess Testing Indicators: - **Complex multi-step workflows** (>5 sequential actions) - **Visual validation requirements** (UI appearance, layout) - **Exploratory testing needs** (finding edge cases) - **Cross-browser/cross-platform testing** - **End-to-end flows spanning multiple pages/services** - **Natural language test specifications in JIRA** - **Need for autonomous test discovery** ### Direct Testing Indicators: - **Simple CRUD operations** - **Unit test level functionality** - **Single endpoint/component testing** - **Performance benchmarks** - **Quick smoke tests** - **CI/CD pipeline tests** - **Deterministic test cases** ## Step 4: Execute Appropriate Testing Based on my analysis, I'll proceed with the detected approach: ### If UI Testing Detected: → **Complex UI flows**: Using subprocess with `/project:test-ui-subprocess` → **Simple UI checks**: Using direct Puppeteer with `/project:test-ui` ### If Backend API Testing Detected: → **Complex integrations**: Using subprocess with `/project:test-api-subprocess` → **Simple endpoints**: Using direct pytest with `/project:test-api` ### If MCP Testing Detected: → Always using subprocess testing with `/project:test-mcp` → MCP tools benefit from autonomous agent execution ### If Hybrid Testing Needed: → Will test multiple layers: 1. Backend API (choose direct/subprocess based on complexity) 2. UI interactions (choose direct/subprocess based on flow) 3. MCP tools if applicable (always subprocess) ### If Uncertain: I'll analyze the JIRA issue acceptance criteria more carefully and may ask: - "Would you prefer comprehensive subprocess testing or quick direct testing?" - "Are there specific edge cases or exploratory scenarios to test?" Let me begin the analysis... --- ## 🚀 Next Steps After Testing After tests complete: 1. **If tests pass, complete the feature:** ``` /project:complete-issue [SITE-ALIAS] ``` This will commit your changes, create a PR, and update JIRA 2. **If tests fail:** - Fix the issues identified - Run `/project:test-issue` again - Repeat until tests pass 3. **For specific test types:** - UI issues? Try `/project:test-ui-subprocess` for deeper exploration - API issues? Try `/project:test-api-subprocess` for comprehensive testing - MCP issues? Review the subprocess output carefully 💡 **Tip:** The test command automatically detects the best testing approach based on your changes!