⚡ EXECUTION INSTRUCTIONS FOR CLAUDE

When this command is invoked, YOU (Claude) must execute these steps immediately: This is NOT documentation - these are COMMANDS to execute right now. Use TodoWrite to track progress through multi-phase workflows.

🚨 EXECUTION WORKFLOW

Phase 1: Intelligent Test Selection (Default Behavior)

Action Steps:

1. PR Change Analysis:

   • Get current branch: git branch --show-current
   • Find changed files: git diff --name-only origin/main...HEAD
   • Map changed files to relevant test files using intelligent heuristics
   • Select subset of tests that should run based on changes

2. Test Mapping Logic:

   • Python file changes → corresponding test files
   • Frontend changes → frontend/UI test files
   • Config changes → integration test files
   • If >20 files changed → fall back to full suite

Phase 2: Test Execution with Modifiers

Action Steps:

1. Parse Command Modifiers:

   • Extract test pattern from command (e.g. testsABC from /test testsABC)
   • Support flags: --full, --integration, --coverage
   • Default to intelligent selection if no modifiers

2. Execute Tests with Output Logging:

   • Run ./run_tests.sh with appropriate flags
   • Output automatically saved to /tmp/{branch_name}/run_tests_{timestamp}.txt
   • Show real-time output to user
   • Parse results for pass/fail analysis

Phase 3: Pattern Matching (When Pattern Specified)

Action Steps:

1. Test File Discovery:

   • Find all test files matching pattern: find . -name "*{pattern}*" -path "*/test*" -name "test_*.py"
   • Validate test files exist and are executable
   • Run specific test subset

2. Execution Command Building:

   • Build command: ./run_tests.sh {matched_files}
   • Handle multiple matches appropriately
   • Provide feedback on which tests are being run

Phase 4: Results Analysis & CI Check

Action Steps:

1. Local Test Analysis:

   • Parse test output for pass/fail counts
   • Identify failing tests and root causes
   • Fix any failing tests immediately

2. GitHub CI Status Check:

   • Check current PR/branch status with gh pr checks [PR#]
   • If GitHub tests failing, download logs and fix issues
   • Verify GitHub tests pass after fixes
   • Commands: gh pr checks, gh run view --log-failed

Phase 5: Completion Criteria

Action Steps:

1. All selected tests pass (local)
2. All GitHub CI checks pass
3. Never dismiss failing tests as "minor"
4. Debug root causes of failures
5. Both local and CI must be green before completing
6. Provide summary of test execution and results

📋 REFERENCE DOCUMENTATION

Enhanced Test Command

Purpose: Intelligent test execution with PR-based test selection and modifiers

Action: Execute tests intelligently based on PR changes or with custom modifiers

Usage:

• /test - Run tests for files changed in current PR (default)
• /test [pattern] - Run tests matching pattern (e.g. /test testsABC)
• /test --full - Run complete test suite
• /test --integration - Include integration tests

Implementation:

Smart Test Selection Examples:

• Changes to $PROJECT_ROOT/main.py → Run test_main*.py files
• Changes to $PROJECT_ROOT/gemini_service.py → Run test_gemini*.py files
• Changes to frontend files → Run frontend test suite
• Changes to > 20 files → Full test suite
• No changes detected → Run minimal smoke test suite