webmcp-polyfill
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's core workflow requires fetching documentation from external sources, which may contain adversarial content.
- Ingestion points: The 'Before writing code' section instructs the agent to use WebSearch and WebFetch to retrieve README files and API documentation from GitHub and NPM.
- Boundary markers: There are no instructions or delimiters provided to ensure the agent ignores or sanitizes potential instructions embedded within the fetched external data.
- Capability inventory: The skill is granted Bash, Write, Edit, and WebFetch tools. A successful injection could lead to unauthorized file modifications or command execution.
- Sanitization: No validation or sanitization process is defined for the content retrieved from external URLs.
- [EXTERNAL_DOWNLOADS]: The skill guides the agent to locate and install external Node.js packages based on dynamic search results rather than fixed references.
- Evidence: The 'Package Structure' and 'Feature Detection' sections encourage the agent to find the 'latest package names' on NPM. This reliance on search results for identifying dependencies poses a risk of dependency confusion or installing malicious packages masquerading as the intended library.
Audit Metadata