playwright-tutorial-video

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt's capture utilities record typed values into the manifest (typedValue) and the generation steps then bake those values verbatim into generated files (data/flow.ts and scene constants), which would force the agent to output any sensitive inputs exactly as entered.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill's capture/recording flows explicitly instruct Playwright to visit an arbitrary App URL and produce a manifest (public/screenshots//manifest.json) that the agent reads and uses (captions, selectors, element coords, typed values) to decide component mappings, timings, and generate scenes, so untrusted third‑party webpage content can directly influence the agent's decisions and tool use.