ml-model-evaluation
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is composed exclusively of Markdown documentation and templates designed to guide human or agent workflows.
- [NO_CODE]: There are no scripts, binaries, or automated tasks included in the skill files; it functions purely as instructional context.
- [EXTERNAL_DOWNLOADS]: The skill does not reference, fetch, or download any external resources, packages, or code from the internet.
- [DATA_EXFILTRATION]: No network communication patterns, API keys, or sensitive local file system access were identified.
- [PROMPT_INJECTION]: The instructions focus on model evaluation metrics and do not contain patterns designed to override agent safety filters or system prompts.
Audit Metadata