run-smoke-tests — Claude Code skill: Playwright smoke tests

What run-smoke-tests does

The skill walks the agent through one path: build the app, run the smoke suite or a single spec, and on failure read traces and logs to pin down the cause. Then it applies the smallest fix and reruns until the result stops moving.

Three npm commands are baked in. npm run smoketest covers the whole suite, npm run smoketest -- path/to/test.spec.ts targets one file, and npm run smoketest-no-compile -- path/to/test.spec.ts skips the build step when artifacts already exist.

Your project has to define those npm scripts and have Playwright in place, otherwise the steps have nothing to call. The guardrails push deterministic waits and assertions over timeouts, ask for a rerun after a passing fix, and keep tests out of quarantine unless someone asks for it in writing. Expect a results summary, the root cause with its fix, and whatever flake risk is left.