idea-refine
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The script
scripts/idea-refine.shis used for environment initialization. It only creates a directory (docs/ideas) and does not perform any dangerous operations or request sudo privileges. - [EXTERNAL_DOWNLOADS] (SAFE): No external dependencies or remote scripts are downloaded.
- [PROMPT_INJECTION] (LOW): The skill processes local codebase files which presents an indirect prompt injection surface. 1. Ingestion points: Codebase files read via
GlobandReadtools. 2. Boundary markers: Absent. 3. Capability inventory: Local file writing and conversational response. 4. Sanitization: Absent. This surface is expected for the skill's functionality. - [DATA_EXFILTRATION] (SAFE): The skill does not have network access or capabilities to transmit data to external domains.
Audit Metadata