brainstorming
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted project data to create design specs.
- Ingestion points: The skill instructions command the agent to read project files, documentation, and recent commits.
- Boundary markers: No explicit boundary markers or warnings are defined for the file-reading process.
- Capability inventory: The skill utilizes a local server with file system and process launching capabilities, then transitions to implementation skills.
- Sanitization: Prompt-level sanitization is not applied to the project data, although the server does escape HTML for its own UI.
- [COMMAND_EXECUTION]: The server script (scripts/server.cjs) uses Node.js child_process to launch the browser. This is a functional requirement and uses either standard platform launchers or a user-provided command.
- [EXTERNAL_DOWNLOADS]: The visual companion fetches a logo from primeradiant.com and refers to its GitHub repository for documentation.
Audit Metadata