CyanAutomation avatar

judokon-test-author

Writes and updates automated tests for JU-DO-KON! using Vitest and Playwright. Use whenever logic, s

提供方 CyanAutomation|开源

Skill Instructions

Inputs / Outputs / Non-goals

  • Inputs: changed logic, affected features, relevant test locations.
  • Outputs: updated tests, coverage notes, targeted test commands.
  • Non-goals: unnecessary full-suite runs or DOM-mutation testing.

Trigger conditions

Use this skill when prompts include or imply:

  • Adding or modifying battle logic.
  • Changing state transitions or gameplay rules.
  • Introducing feature flags or bug fixes.

Mandatory rules

  • No logic change without tests.
  • Test behavior through public APIs/user interactions; avoid direct DOM mutation.
  • Cover state transitions and feature-flag on/off behavior.
  • Prefer targeted Vitest/Playwright runs for changed areas.
  • Keep test logs clean (no unsilenced console.warn/error).

Validation checklist

  • Happy-path and edge-case coverage added for new behavior.
  • Targeted test commands executed (e.g., npx vitest run tests/<path>.test.js).
  • Relevant battle test scripts used when applicable (npm run test:battles:classic, npm run test:battles:cli).
  • Console discipline and timer determinism expectations met.

Expected output format

  • List of new/updated test files with behavior covered.
  • Exact test commands run and outcomes.
  • Coverage gap notes and recommended follow-ups.

Failure/stop conditions

  • Stop when required behavior cannot be tested deterministically.
  • Stop when requested test strategy depends on prohibited anti-patterns.
judokon-test-author - 适用于 Claude Code 与 Cursor 的 AI 智能体 Skill | Agent Skills