capture-project-vision
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and processes local file content (e.g., README.md, CONTEXT.md), which constitutes an untrusted data surface. To mitigate risks, the instruction set mandates an explicit 'proposal and approval' phase where the user must review suggested edits before the agent writes to the filesystem.
- [EXTERNAL_DOWNLOADS]: The documentation references its origin in a Vercel Labs GitHub repository. This is a reference to a trusted organization and does not involve automated execution of remote code.
Audit Metadata