slime-rl-training
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides comprehensive and legitimate technical documentation for the slime framework, following industry-standard practices for reinforcement learning training without any detected malicious patterns.\n- [EXTERNAL_DOWNLOADS]: The instructions reference downloading source code from the official THUDM GitHub repository and utilizing Docker Hub for environment containerization, both of which are standard and recognized sources in the AI research community.\n- [INDIRECT_PROMPT_INJECTION]: The skill describes ingesting training data from external JSONL files; while this represents a data ingestion surface, it is a core functional requirement of the training process and is documented as such.
Audit Metadata