unsloth
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The entire set of files contains only markdown documentation and python code fragments showcasing standard configuration interfaces.
- [SAFE]: Instructions cover genuine workflows for supervised fine-tuning (SFT) and reinforcement learning (GRPO, DPO).
- [SAFE]: Package recommendations are un-obfuscated and use established tools from standard repositories (
unsloth,vllm,transformers,trl,peft,datasets,sentence-transformers). - [SAFE]: Configuration scripts target legitimate engineering backends like Modal, Hugging Face Hub, and llama.cpp without credential risks or unauthorized telemetry endpoints.
Audit Metadata