axolotl
Pass
Audited by Gen Agent Trust Hub on Jun 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [SAFE]: The skill provides documentation and reference materials for the Axolotl LLM fine-tuning framework. It does not include any executable scripts or malicious instructions.
- [EXTERNAL_DOWNLOADS]: Mentions downloading configuration, model weights, and datasets from well-known services such as Hugging Face Hub and GitHub. Specifically, it documents the
fetch_from_githubutility and thetrust_remote_codeconfiguration option used in Hugging Face Transformers. - [COMMAND_EXECUTION]: Provides reference shell commands for training, preprocessing, and inference using the Axolotl CLI (e.g.,
axolotl train,axolotl inference). These are instructional and essential to the tool's primary purpose. - [DATA_EXFILTRATION]: Documents configuration for remote storage providers (AWS S3, Google Cloud Storage, Azure, OCI) for dataset loading. The instructions correctly advise using environment variables for credentials rather than hardcoding sensitive data, adhering to security best practices.
Audit Metadata