# Tdd > Strict test-driven development workflow for implementing features Use when this capability is needed. - Skill: `tomevault-io/tdd-3` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/tdd-3` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/tdd-3/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Productivity - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/tdd-3 --- Implement the following feature using strict TDD: $ARGUMENTS Think about how to break down the feature into small, testable units (if applicable). Then, for each unit, follow the strict TDD workflow below: 1. Write behavior-oriented tests based on expected input/output pairs - test WHAT the code does, not HOW it does it 2. Do NOT create mock implementations, unless strictly necessary - write real tests that will fail 3. Run `uv run pytest -v` only on the new tests to confirm they fail 4. Write code that passes the tests 5. Do NOT modify the tests during implementation, unless strictly necessary - tests should verify behavior, not implementation details 6. Run `uv run pytest -q --tb=short` to verify tests pass (quiet mode, concise failures). If many tests fail, use `-x` to stop on first failure and fix incrementally 7. Commit the implementation Once all units are complete and passing, use `gh pr create` to open a PR with test and implementation commits. IMPORTANT: Tests first, implementation second. Never modify tests during step 6. Tests should verify behavior (inputs/outputs), not implementation details. --- > Converted and distributed by [TomeVault](https://tomevault.io/claim/cverrier) — claim your Tome and manage your conversions.