testing-with-playwright
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions and delegation patterns allow the agent to process data from the web application's environment, which constitutes an ingestion point for untrusted content.
- Ingestion points: Data is retrieved from browser console logs, network requests, and accessibility snapshots at http://localhost:5173 as specified in SKILL.md.
- Boundary markers: The delegation prompt template lacks explicit delimiters or instructions to ignore instructions embedded in application data.
- Capability inventory: The skill uses Playwright MCP to interact with the browser (Navigate, Click, Type) and capture snapshots.
- Sanitization: No sanitization or validation of retrieved browser data is documented.
Audit Metadata