| name | test-strategy |
| display_name | Test Strategy |
| description | Designs a focused test plan for a feature, bug fix, refactor, or release by choosing the right unit, integration, end-to-end, and regression tests. |
| category | testing |
| version | 1.0.0 |
| tags | ["testing","coverage","strategy","regression","quality"] |
| providers | ["claude-code","codex","copilot","openclaw","hermes-agent"] |
| author | agent-skills contributors |
Test Strategy
Use this skill when a user asks what to test, how much to test, or how to validate a change before release.
Inputs
- Feature, bug fix, refactor, pull request, or release scope
- Existing test framework and commands
- Known risks, production incidents, or edge cases
Steps
- Identify behaviors that must remain true, not just lines of code that changed.
- Map risks to test levels: unit for pure logic, integration for contracts and data flow, end-to-end for critical user journeys, regression tests for bugs.
- Prioritize tests by impact, likelihood, and cost to maintain.
- Include negative cases, boundary cases, permissions, failure modes, and compatibility cases.
- Recommend test data and mocks only where they preserve realism.
- Define the smallest useful test suite for this change and the broader suite for release confidence.
- Provide commands to run if they can be inferred from the repo.
Output
Return:
- Risk-based test matrix
- Recommended new or updated tests
- Existing tests to run
- Gaps that require manual QA or instrumentation
- Confidence assessment
Guidelines
- Avoid blanket "add more tests" advice.
- Prefer tests that fail for the right reason.
- Keep tests aligned with user-visible behavior and stable contracts.
