organise-agent-skills
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to perform file system operations including directory listing, file reading, and file modification within the agent's configuration paths (e.g.,
~/.gemini/config/). - Evidence: Instructions include listing directories under
~/.gemini/config/skills/, readingskills.json, and updating documentation files likeGEMINI.mdandREADME.md. - [COMMAND_EXECUTION]: The skill allows for the deletion of obsolete skill directories.
- Mitigation: The skill includes a 'Mandatory User Confirmation Guardrail' that strictly forbids deletion or configuration mutation without explicit user approval and a presented audit of proposed changes.
- [SAFE]: No obfuscation, remote code execution, or data exfiltration patterns were detected. The skill's behavior is consistent with its stated purpose of optimizing the agent's context window through skill management.
Audit Metadata