Write Tests

功能

• Adds tests for uncommitted and untracked files.
• Covers latest commit if all changes are already committed.
• Orchestrates specialized review and development agents for test creation.
• Analyzes code complexity to determine agent orchestration vs. direct test writing.
• Verifies test coverage and iterates until critical logic is covered.

使用场景

• When implementing new features and needing to ensure they are adequately tested.
• After refactoring existing code to add tests for the modified logic.
• To improve overall test coverage of a codebase incrementally.
• As part of a CI/CD pipeline to automatically generate tests for developer changes.

非目标

• Modifying existing tests.
• Writing tests for code changes that are not local.
• Achieving 100% line coverage for every single line of code.
• Replacing human judgment entirely; it's an augmentation tool.

工作流

1. Read sadd skill if available.
2. Discover test infrastructure (read README/package.json, identify test commands, understand conventions).
3. Run all tests to establish baseline.
4. Analyze local changes: git status, filter non-code files, launch agents per file for complexity analysis.
5. If single file and simple change: Read TDD skill, understand file logic, review existing tests, add cases, create tests, run tests, iterate.
6. If multiple files or complex change: Launch `review:test-coverage-reviewer` agents per file (parallel) for coverage reports, then launch `developer` agents per file for test creation.
7. Verify coverage using `review:test-coverage-reviewer` agents, iterating if gaps are found.
8. Final verification: run full test suite, generate coverage report, verify no regressions.

实践

• Test-Driven Development
• Code Quality
• Agent Orchestration

先决条件

• Git installed and configured.
• Access to a suitable AI model for agent execution (e.g., Sonnet or Opus models mentioned in SKILL.md).