npmjs
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests search suggestions and metadata from npmjs.com, which is an untrusted external source. An attacker could potentially influence agent behavior by publishing packages with malicious descriptions.\n
- Ingestion points:
npmjs_com__get_search_suggestions(SKILL.md).\n - Boundary markers: None specified in the instructions.\n
- Capability inventory: Remote task execution, scraping, and cloud browser interaction via vendor tools.\n
- Sanitization: Not described in the skill content.\n- [EXTERNAL_DOWNLOADS]: The skill is configured to connect to a remote MCP server at
https://unbrowse.ai/mcp/npmjs.com. This is a vendor-owned resource used to provide the skill's functionality.\n- [DYNAMIC_EXECUTION]: The skill provides tools for interacting with a cloud browser (unbrowse_browse_open,unbrowse_browse_act) which can be used to record and compile new tools at runtime on the vendor's infrastructure.
Audit Metadata