| name | qa-testing |
| description | QA and testing knowledge. Use when testing an implementation end-to-end — building a test plan, validating APIs with curl, driving the UI with Playwright, comparing against Figma, and documenting pass/fail with evidence. |
| triggers | ["qa","test","testing","validate","playwright","e2e","test plan"] |
QA Testing
Use this skill when validating an implementation across any affected layer (backend APIs and/or frontend UI).
Approach
- Identify all test cases from the requirements (happy path, edge cases, errors, UI flows, interactions)
- Start the affected services if not running
- Test backend cases using curl, scripts, or API calls
- Test frontend cases using Playwright (browser automation)
- Compare UI with Figma designs if available
- Document pass/fail with evidence (screenshots when relevant)
Tools
- curl for API validation
- database MCPs to verify data state after operations
- Playwright MCP for browser-based testing
- Figma MCP to compare implementation with design
Security
- Never access tokens or credentials directly from the database
- Use environment variables for test user tokens
- Never include tokens or secrets in documents or output
Documenting results
Separate results by repository:
### Repository: api
#### Test: Create user profile
- Status: PASS
- Evidence: 201 response with expected payload
### Repository: frontend
#### Test: Form validation on submit
- Status: FAIL
- Evidence: Email field accepts invalid format
- Expected: Inline validation error
If you need the user to do something (DB changes, restart services), document it clearly, continue with independent tests, and mark dependent tests as waiting.
