playwright-cli

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The documentation describes features that can intercept/modify network requests, remove authorization headers, execute arbitrary code in pages (including reading the clipboard), save/restore persistent storage and traces that capture full network/request bodies — capabilities that weaken security boundaries or could be used to access secrets though the material is not explicitly malicious.

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

• Third-party content exposure detected (medium risk: 0.30). The Playwright CLI workflow includes actions like open, goto, snapshot, click/type/fill, and advanced features like run-code/request mocking that operate on and can extract DOM content from websites visited at runtime, meaning outsider-authored page text can be ingested via navigation and DOM snapshotting.