Implement features using TDD and specific Yuzu validation protocols.

Install

mkdir -p .claude/skills/tdd-tr3kkr && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/13469" && unzip -o skill.zip -d .claude/skills/tdd-tr3kkr && rm skill.zip

Installs to .claude/skills/tdd-tr3kkr

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Apply Yuzu-specific test-driven development using tracer-bullet red-green-refactor cycles and the repo's Meson, gateway, proto, and ABI validation rules. Use when the user says `/tdd`, asks for TDD, red-green-refactor, or wants a feature/fix built test-first in Yuzu.
267 chars✓ has a “when” triggerlonger than Claude Code's old 250-char listing cap (fine on current versions)
Advanced

Key capabilities

  • Add a single failing test for one observable behavior
  • Implement minimal code to pass the failing test
  • Verify test execution using `meson test` or gateway commands
  • Repeat the red-green cycle for subsequent behaviors
  • Refactor code only when all tests are green
  • Use Yuzu-specific test hygiene practices for temporary files and isolation

How it works

The skill guides through a test-driven development cycle, focusing on public interfaces and Yuzu-specific build and validation tools.

Inputs & outputs

You give it
A new feature or bug to be implemented test-first
You get back
Code changes with passing tests, and targeted validation output summarized

When to use tdd

  • Implementing new features
  • Fixing bugs via TDD
  • Running Yuzu integration tests

About this skill

TDD

Use behavior-first tests through public interfaces. Prefer integration-style seams that exercise real stores, REST handlers, gateway paths, plugin boundaries, or command dispatch contracts. Avoid tests coupled to private implementation.

Before Red

  • Read CONTEXT.md and relevant ADRs in docs/adr/.
  • Identify the public interface and the project term for the behavior.
  • List the smallest vertical behavior slice that proves value.
  • Choose the narrowest validation path through $yuzu-build.
  • Invoke $yuzu-proto, $yuzu-plugin-abi, $yuzu-meson, or $yuzu-windows-msvc if the slice crosses those boundaries.

Cycle

  1. Red: add one failing test for one observable behavior.
  2. Green: implement only enough code to pass that test.
  3. Verify: run the targeted suite with meson test -C <builddir> --suite <suite> --print-errorlogs or the relevant gateway command.
  4. Repeat for the next behavior.
  5. Refactor only while green.

Do not write all tests first. Use tracer bullets: one behavior, one test, one implementation.

Yuzu Test Hygiene

  • Use yuzu::test::unique_temp_path(prefix) or TempDbFile for C++ test temp files and SQLite DBs.
  • Avoid hardcoded shared paths and clock/thread-hash uniqueness.
  • For gateway EUnit, pass --dir apps/yuzu_gw/test.
  • For direct Erlang isolation, remove the relevant _build/test tree or run in a fresh BEAM.
  • For proto changes, include buf lint proto and buf breaking proto --against '.git#subdir=proto,ref=origin/main' when available.
  • For plugin ABI changes, add descriptor/loader coverage and preserve stable C ABI compatibility.

Done

Finish with the tests added, implementation green, targeted validation output summarized, and any broader /test --quick or governance follow-up called out.

When not to use it

  • The user wants to write all tests before any implementation
  • The project is not Yuzu-specific or does not use Meson, gateway, proto, or ABI validation rules

Limitations

  • It is specific to Yuzu projects
  • It requires familiarity with Meson, gateway, proto, and ABI validation rules
  • It emphasizes integration-style seams over unit tests

How it compares

This workflow enforces a strict red-green-refactor cycle with Yuzu-specific tools and hygiene, ensuring behavior-first development and integration seam testing.

Compared to similar skills

tdd side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
tdd (this skill)02moNo flagsAdvanced
code-coverage-with-gcov154moReviewIntermediate
unity-mcp-orchestrator174moNo flagsIntermediate
add-uint-support189moNo flagsIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry