jarvis-setup
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies ingestion points for untrusted data (e.g., Notion pages via MCP, repository files, screen content) and documents a built-in security mechanism (TAINTING_TOOLS) designed to prevent acting on potentially malicious instructions found in that data.
- [EXTERNAL_DOWNLOADS]: Documents requirements for official packages such as @anthropic-ai/claude-code and @notionhq/notion-mcp-server, which are standard tools from trusted providers.
- [COMMAND_EXECUTION]: Mentions the use of commands like npx, python, and claude as part of the setup and runtime environment. These are expected for a development tool of this nature.
Audit Metadata