
from dotfiles33
Guides an agent through a strict red–green–refactor TDD cycle: write a failing test, implement the minimal change, and refactor with verification.
This skill embeds a disciplined Test-Driven Development workflow into agent work: identify the smallest observable behavior change, write a failing test, make the minimal production change to pass, then refactor while keeping tests green. It forces incremental, verifiable progress so agents avoid speculative or broad changes without test coverage.
Use this skill when implementing new features, fixing reproducible bugs, adding regression coverage, or changing public business logic where tests can verify behavior. Do NOT use for documentation-only edits, purely mechanical renames, or when no test harness exists.
Best suited to coding-capable agents that can run project tests and edit code (Copilot-style agents, Codex, GitHub Copilot Chat, or similar). The skill assumes access to the repository's test tooling and the ability to run test commands and edit files.
Pure instructional SKILL.md with no bundled scripts. Guides agents through a strict red-green-refactor TDD cycle with clear workflow steps, test selection heuristics, anti-patterns, and an output template. Well-written and unambiguous — a coding agent could follow this immediately. No security concerns whatsoever as there are no executable components.
Clean instructional skill. Would benefit from a references/ section with common test framework command patterns (jest, pytest, etc.) to reduce the 'ask one question' friction. Output template is a nice touch for status updates.