Write tests
Skill Karthick-Ramachandran/persist-os/.agents/skills/write-tests
Local-first CLI that gives your repo durable, AI-ready memory - features, ADRs, modules, architecture, standards & tests and a deterministic doctor that validates health, evidence & drift.
npx -y skills add Karthick-Ramachandran/persist-os --skill write-testsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 8 stars8 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.
SKILL.md
1.8 KB, as published. Nobody here has run it
Skill: Write Tests
Purpose
Create professional tests that prove important behavior, not random happy paths.
Inputs
- Feature PRD or task.
- Acceptance criteria.
- Changed files or planned modules.
- Known risks.
Required Reading
docs/50-quality/TESTING_STRATEGY.mddocs/50-quality/QUALITY_GATES.mddocs/20-security/SECURITY_MODEL.mddocs/60-engineering/ENGINEERING_STANDARDS.md- Relevant feature
TEST_PLAN.md - Relevant module
TEST_PLAN.md
Output Files
- Test files under the appropriate
tests/area. - Updated
TEST_PLAN.mdwhen test strategy changes. - Completion evidence listing commands and results.
Process
- Map acceptance criteria to test cases.
- Add risk-based tests for unsafe paths, overwrites, symlinks, config validation, generated output, and CLI behavior as relevant.
- Prefer unit tests for pure logic and integration tests for command behavior.
- Add golden tests for generated docs and templates.
- Name tests by behavior.
- Document skipped tests and remaining risk.
Stop Conditions
Stop and request human decision if:
- Requirements are not testable.
- The requested test would require runtime network, telemetry, cloud, MCP, or AI API behavior without review.
- Security-sensitive behavior lacks a documented expected result.
- The test approach conflicts with engineering standards.
Quality Bar
- Tests derive from requirements and risk.
- Security invariants are covered.
- Test names describe behavior.
- Engineering standards for evidence and skipped checks are followed.
- The completion report includes commands and results.