Principles
- Test behavior, not implementation details
- Prefer integration tests over unit tests (Testing Trophy)
- Arrange-Act-Assert (AAA) pattern in every test
- Tests must be independent — no shared mutable state
- Keep tests small and focused — one behavior per test
- Name tests to describe the behavior being verified
- Optimize for confidence, not coverage percentage
- Don't chase 100% coverage — test what matters
Rules
See rules index for detailed patterns.
Examples
Positive Trigger
User: "Define integration-vs-unit test boundaries and mocking strategy."
Expected behavior: Use platform-testing guidance, follow its workflow, and return actionable output.
Non-Trigger
User: "Create a tRPC router for billing procedures."
Expected behavior: Do not prioritize platform-testing; choose a more relevant skill or proceed without it.
Troubleshooting
Skill Does Not Trigger
- Error: The skill is not selected when expected.
- Cause: Request wording does not clearly match the description trigger conditions.
- Solution: Rephrase with explicit domain/task keywords from the description and retry.
Guidance Conflicts With Another Skill
- Error: Instructions from multiple skills conflict in one task.
- Cause: Overlapping scope across loaded skills.
- Solution: State which skill is authoritative for the current step and apply that workflow first.
Output Is Too Generic
- Error: Result lacks concrete, actionable detail.
- Cause: Task input omitted context, constraints, or target format.
- Solution: Add specific constraints (environment, scope, format, success criteria) and rerun.
Workflow
- Identify whether the request clearly matches
platform-testingscope and triggers. - Apply the skill rules and referenced guidance to produce a concrete result.
- Validate output quality against constraints; if gaps remain, refine once with explicit assumptions.