Test driven development

Skill g-shevchenko/agentic-quality-skills/skills/test-driven-development

Production-grade quality skills for AI coding agents: red-first TDD, quality gates, and golden benchmark uplift loops.

Install

npx -y skills add g-shevchenko/agentic-quality-skills --skill test-driven-development

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use when implementing a behavior change, fixing a bug, writing tests, or when the user says use test driven development, use TDD, test first, failing test first, verify red, iron law, fail-to-pass, lock tests, test immutability, no test relaxation, or frontier-only TDD. Requires a verified failing test before implementation, forbids weakening existing tests, and routes strict TDD implementers to frontier-class models.

SKILL.md

2.0 KB, as published. Nobody here has run it

Test Driven Development

Use this skill to keep agent-written code honest: prove the missing behavior with a failing test, then implement the smallest code change that turns the test green.

Workflow

Clarify the behavior being changed in one acceptance criterion.
Write or select the test that should fail before the fix.
Run the test before implementation.
Verify RED:
- the test fails;
- the failure is an assertion or expected exception;
- the failure message matches the missing behavior;
- the failure is not a broken import, syntax error, missing symbol, or setup error.
Implement the smallest production change.
Re-run the targeted test, then the nearest relevant suite.
Report the red proof and green proof.

For detailed rules, read references/tdd-rules.md.

Test Immutability

Existing tests are locked. Add tests and assertions; do not weaken, delete, skip, loosen, or rewrite existing expected behavior during a fix loop.

If an existing test appears genuinely wrong, stop and treat it as a spec change. Get approval or a fresh verifier before changing it.

Model Routing

Strict TDD is not a cheap-model task. If an implementer must follow a verified failing test contract, use a frontier-class model. Weaker models can be used for read-only verification, formatting, or non-TDD utility work.

Done Criteria

Do not say the work is done unless the answer includes:

test file and test name;
command that proved RED;
command that proved GREEN;
any exception used, if strict TDD was not appropriate.