local-llm-expert
Pass
Audited by Gen Agent Trust Hub on May 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is purely instructional, focusing on technical advice for local AI deployment and hardware optimization. It does not contain executable scripts or hidden logic.
- [PROMPT_INJECTION]: No instructional overrides, jailbreak patterns, or attempts to bypass safety filters were found.
- [DATA_EXFILTRATION]: No access to sensitive local files (e.g., credentials, SSH keys) or network exfiltration patterns were identified.
- [REMOTE_CODE_EXECUTION]: While the skill suggests command-line tools for local model inference, it does not include instructions for downloading or executing unverified remote code.
Audit Metadata