brainstorm

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download its core components (SKILL.md and gemini.py) from the author's public GitHub repository using curl for installation purposes.
  • [COMMAND_EXECUTION]: The skill's instructions involve executing shell commands to locate its script, check environment variables, and run a local Python wrapper for the Gemini API. It also handles the installation of the necessary Python library (google-genai) if not present.
  • [CREDENTIALS_UNSAFE]: The skill requires a Google API key and provides standard guidance on storing it in a .env file or environment variable, which is a safe practice for API key management.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided context to facilitate brainstorming. It interpolates this context into prompts sent to the Gemini model, which is the intended functional behavior for this type of agent skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 03:11 PM
Security Audit — agent-trust-hub — brainstorm