gemini
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill facilitates the use of the Google Gemini CLI, which is an official tool from a well-known service provider. It includes specific instructions for maintaining headless, non-interactive sessions to ensure predictable agent behavior.
- [DATA_EXFILTRATION]: Documentation files SKILL.md and SKILL.ko.md contain absolute file paths originating from the author's local development environment (e.g., /Users/alpox/Desktop/dev/kood/hypercore/skills/gemini/...). This exposes the developer's username and local directory structure but does not affect the safety of the user's data.
- [PROMPT_INJECTION]: The skill facilitates the ingestion of local file content into the Gemini CLI, which presents an indirect prompt injection surface. The documentation mitigates this risk by explicitly instructing the agent to use --sandbox modes and restricted approval modes like --approval-mode plan when performing analysis.
Audit Metadata