testing
the proof seeker
testing work starts from what must be true, not from which tool to run: the claim under test picks the suite, and the suite picks the tool. a green run that misses the risky paths proves little; a flaky suite is a bug, not background noise.
tests the contractflakes are bugsred before green
Use for any testing work -- running unit, integration, e2e, UI, and visual regression tests with coverage analysis and QA reports, or web test automation with Playwright, Vitest, and k6 (e2e/unit/integration/load/security/visual/a11y testing, flakiness, Core Web Vitals, mobile gestures, cross-browser).
methodology
- classify the task: running or designing validation suites, or web test automation (browser, load, visual, a11y).
- read the matching guide below before writing or running tests.
- run the smallest meaningful check for the changed behavior and any required project gate; avoid tests that merely mirror implementation or impose a generic coverage target.
- report observed output and proof boundary; chase flakiness to a root cause instead of retrying it away.
- use the trained decisions when bundled examples assume unavailable tools or fixed thresholds.
contents
- trainedlearned layer -- read first
- testrunning unit, integration, e2e, or UI tests; coverage analysis; build verification; visual regression; QA reports; test strategy
- web-testingweb test automation with Playwright, Vitest, or k6 -- e2e/load/security/visual/a11y, flakiness, Core Web Vitals, mobile gestures, cross-browser