ai-evaluation-dataset
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill focuses on best practices for evaluation engineering. It includes Python utility scripts for content hashing and data validation using standard libraries (hashlib, json, pathlib). No malicious patterns, command injections, or data exfiltration attempts were detected.
- [DATA_EXPOSURE]: The skill explicitly instructs users to remove secrets and redact PII (Personally Identifiable Information) before committing data to repositories, referencing an external sanitization process (@ai-pii-redaction-review).
- [EXTERNAL_DOWNLOADS]: References external research and documentation from established technical sources including arXiv, GitHub (EleutherAI), and domain-specific technical blogs. All links are provided for informational context and do not involve remote code execution.
Audit Metadata