| name | testing |
| description | Write, fix, and run unit, integration, regression, browser, and end-to-end tests. Use when proving behavior, debugging failures, or improving coverage. |
Testing
Use this skill for automated tests, coverage improvements, regression tests, and end-to-end or browser verification.
Workflow
- Identify the behavior to prove.
- Prefer writing or updating the test before changing implementation when practical.
- Discover the repo's test commands from scripts, docs, or existing CI config, then run the narrowest relevant test first.
- If browser behavior matters and the user enabled Playwright MCP, use
mcp-playwright.
- If the failure pattern or test design is still unclear and the user enabled Open Bridge, use
mcp-open-bridge for a second-pass test or diagnosis review.
- After the focused test passes, run broader relevant verification. If a failure is unrelated or environment-caused, record it separately instead of expanding the change.
Relevant MCP Skills
mcp-context7 for framework-specific testing APIs
mcp-playwright for browser flows, snapshots, and UI checks
mcp-github for CI failures or test-related PR context
mcp-open-bridge for flaky-test hypothesis review and alternative regression-test ideas
Output
Include:
- Tests added or updated
- Focused verification run
- Broader verification run
- Remaining gaps