quickbrain

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a custom Node.js web server (server.cjs) and shell scripts to host a 'Visual Companion' interface. This server allows the agent to display HTML-based wireframes and diagrams. The server binds to localhost by default and includes logic to automatically shut down when the parent process exits or after a period of inactivity.
  • [INDIRECT_PROMPT_INJECTION]: The workflow processes user requirements to generate plans executed by subagents. It mitigates injection risks through a 'Plan Fidelity Rule' and mandatory two-stage reviews by independent subagents before changes are finalized.
  • [SAFE]: The skill follows security best practices, including using standard libraries for its server to avoid supply chain risks and implementing path sanitization to prevent directory traversal.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 12:42 AM