workflow engine that enforces strict separation of structural and functional changes for maximum review efficiency. Use when the user requests implementation of new features, bug fixes, performance improvements, or refactoring. Triggers on requests like "implement [feature]", "fix [bug]", "optimize [performance]", "add [functionality]", or when the user provides specifications via file or description. Also triggers when user explicitly invokes "/autopilot". --- ~/.claude/skills/autopilot/SKILL.md
and Test Foundation (Test-First) **Objective:** Understand requirements and ensure safe refactoring through adequate test coverage. **Actions:** 1. Analyze the requirement specification (from user message or provided file) 2. Identify the code areas that will be modified 3. Read existing tests for those areas 4. Identify missing test coverage (edge cases, error paths, integration points) 5. If tests are insufficient, add tests to cover existing behavior BEFORE any changes **Test Execution:** Run tests after adding new test coverage - Command: project-specific test command (e.g. `npm test`, `./gradlew test`) - Must pass before proceeding **Commit:** Use `/commit` skill for test additions ~/.claude/skills/autopilot/SKILL.md
**Objective:** Implement the actual behavioral changes using Test-Driven Development. **TDD Cycle:** Use `/tdd` skill to iterate through the Red-Green-Refactor process for each small unit of logic. **IMPORTANT:** Create a separate commit after completing EACH TDD cycle (each Red-Green-Refactor cycle). Do NOT batch multiple behaviors into one commit. **What happens in this phase:** - Implement new business logic - Add new features - Fix bugs - Connect new structures to actual functionality - Modify algorithms and data transformations - Add validation and error handling **Test Execution:** Run tests after each TDD cycle - New tests should pass - All existing tests must still pass ~/.claude/skills/autopilot/SKILL.md
all requirements are met and the system is stable. **Actions:** 1. Run complete test suite 2. Verify all tests pass 3. Confirm all requirements from Phase 1 are satisfied 4. Check for any edge cases or error scenarios **Test Execution:** Full test suite including: - Unit tests - Integration tests - E2E test (if applicable) **Commit:** Usually not needed unless final adjustments are required **Output:** Fully implemented, tested, and verified feature/fix ~/.claude/skills/autopilot/SKILL.md