Brummer MCP Server

# Test Management Feature Summary ## Overview The Test Management System is the **highest priority feature** in Brummer's roadmap, designed to provide comprehensive testing capabilities with minimal context bloat for AI agents. ## Key Design Goals ### 1. Minimal Context for Success ✅ ``` ✅ 23 tests passed (1.2s) ``` - Single line for all passing tests - No detailed output cluttering AI context - Only timing and count information ### 2. Rich Context for Failures ❌ ``` ❌ user/profile_test.go (line 23) Expected: "john@example.com" Actual: "" Stack: TestUserProfile -> validateEmail -> isEmpty ``` - Complete error information - File locations and line numbers - Stack traces and failure context - Suggested fixes when possible ### 3. Agent-Optimized Events 🤖 - Automatic notifications on test failures - Context-aware filtering (only tests related to recent changes) - Smart relevance scoring for AI agents ## User Interface ### TUI Test Tab - **Hotkey**: `t` to switch to test view - **Run Tests**: `r` for all tests, `R` for pattern-based - **Watch Mode**: `w` for continuous testing - **Results View**: Progressive disclosure (summary → details) ### Key Features - **Real-time Results**: Live updates during test execution - **Coverage Integration**: Optional code coverage display - **Framework Detection**: Automatic test runner detection - **Multi-language Support**: Jest, Go test, pytest, and more ## MCP Integration ### Core Tools ```yaml test_run: # Execute tests with filtering test_status: # Get current test status test_failures: # Get detailed failure info test_coverage: # Code coverage data test_watch: # Control watch mode test_clear: # Clear test history ``` ### AI-Optimized Responses - **Passing Tests**: Minimal JSON with counts only - **Failing Tests**: Rich context with error details - **Smart Filtering**: Only relevant tests for recent changes ## Technical Architecture ### Component Structure ``` /internal/testing/ ├── manager.go # Test orchestration ├── runner.go # Framework-specific runners ├── parser.go # Output parsing ├── watcher.go # File system watching └── results.go # Result storage /internal/tui/ └── test_view.go # TUI interface /internal/mcp/ └── test_tools.go # MCP tools ``` ### Framework Support - **JavaScript**: Jest, Vitest, Mocha, Playwright - **Go**: go test, testify, ginkgo - **Python**: pytest, unittest - **Rust**: cargo test, criterion ## Implementation Timeline ### Phase 1A: Core Infrastructure (Week 1-2) - [ ] Test runner interface and manager - [ ] Basic framework detection (Jest, Go test) - [ ] Test result parsing and storage - [ ] MCP tools foundation ### Phase 1B: TUI Integration (Week 3-4) - [ ] Test view implementation - [ ] Hotkey bindings and navigation - [ ] Real-time result display - [ ] Progressive disclosure UI ### Phase 1C: Advanced Features (Week 5-6) - [ ] Watch mode implementation - [ ] Coverage integration - [ ] Additional framework support - [ ] AI event optimization ## Success Metrics ### Developer Experience - **Test Feedback Time**: < 5 seconds for simple tests - **Context Switching**: Minimal tool switching during development - **Coverage Visibility**: Easy access to coverage data ### AI Agent Effectiveness - **Context Efficiency**: 90% reduction in irrelevant test output - **Failure Resolution**: AI can resolve 70% of test failures with provided context - **Smart Filtering**: Only show tests related to recent file changes ### System Performance - **Response Time**: Sub-second UI updates - **Memory Usage**: Efficient result storage and cleanup - **Watch Mode**: Responsive file change detection ## Configuration ### Test Settings ```toml [testing] timeout = "5m" max_history = 100 default_coverage = false watch_debounce = "500ms" [testing.jest] coverage_threshold = 80 parallel = true [testing.go] race_detection = true coverage_profile = "coverage.out" ``` ## Future Enhancements ### AI-Powered Features (Phase 1D) - **Intelligent Test Generation**: Generate tests from code analysis - **Flaky Test Detection**: Identify unreliable tests - **Performance Regression Detection**: Alert on timing changes - **Smart Test Selection**: Run only tests likely to fail ### Advanced Integrations (Later Phases) - **CI/CD Integration**: Import remote test results - **Code Review Integration**: Show test status in PRs - **Team Dashboards**: Aggregate metrics across developers ## Getting Started Once implemented, users will: 1. **Press `t`** to switch to test view 2. **Press `r`** to run tests 3. **See minimal output** for passing tests 4. **Get rich context** for any failures 5. **Use `w`** for continuous testing during development The system will automatically detect test frameworks and provide intelligent defaults, making it zero-configuration for most projects. ## Impact on Development Workflow ### Before Test Management - Manual test execution in separate terminals - Context switching between tools - Verbose output cluttering AI conversations - No integrated coverage visibility ### After Test Management - One-key test execution within Brummer - Minimal context for AI agents (passing tests) - Rich debugging info when needed (failing tests) - Integrated coverage and watch mode - Smart test selection based on file changes This feature represents a significant step toward making Brummer a comprehensive development environment that minimizes context bloat while maximizing debugging effectiveness.