strands
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of official Strands Agents packages from PyPI and NPM (e.g.,
strands-agents,@strands-agents/sdk). These are vendor-owned resources and their use is consistent with the skill's purpose. - [REMOTE_CODE_EXECUTION]: The skill uses
uvxto launch MCP servers, including a vendor-specific documentation server and a well-known AWS documentation tool. The documentation includes explicit safety guidance, instructing the user to verify commands and approve the execution of untrusted servers. - [INDIRECT_PROMPT_INJECTION]: The skill provides templates for web applications (FastAPI, Express, Next.js) that ingest user-provided data through request endpoints (e.g.,
req.body.prompt). Although these agents possess tool-calling capabilities, the skill mitigates risk by instructing the user to implement lifecycle bounds, turn limits, and token limits to prevent autonomous runaway execution.
Audit Metadata