llm-training
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- SAFE (SAFE): No executable code or shell scripts were found in the skill. It consists entirely of Markdown documentation.
- SAFE (SAFE): All external references point to legitimate documentation for well-known industry frameworks like Hugging Face (Accelerate, TRL), Microsoft (DeepSpeed), and Unsloth.
- SAFE (SAFE): No patterns of prompt injection, data exfiltration, or obfuscation were detected. The content matches its stated purpose of providing an LLM training framework comparison.
Audit Metadata