experiment-tracking-swanlab
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references and provides installation instructions for standard, verifiable Python packages including
swanlab,pillow,soundfile,transformers,pytorch-lightning, andfastai. These are well-established tools in the machine learning ecosystem. - [COMMAND_EXECUTION]: The skill includes documented CLI commands for the SwanLab tool, such as
swanlab login,swanlab watch, andswanlab sync. These are used for their intended purpose of managing experiment logs and authentication. - [DATA_EXFILTRATION]: The skill's primary function is experiment tracking, which naturally involves sending training metrics and media to a server. This behavior is documented, expected, and limited to parameters provided by the user (project name, metrics, etc.). It does not attempt to access sensitive system files or credentials outside its scope.
- [CREDENTIALS_SAFE]: Code examples demonstrate secure practices for sensitive data, specifically instructing users to load API keys from environment variables (
os.environ["SWANLAB_API_KEY"]) rather than hardcoding them.
Audit Metadata