llm-obs-experiment-py-bootstrap
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the local shell to run syntax checks on generated files using python -m py_compile and a JSON-parsing Python one-liner for Jupyter notebooks.\n- [EXTERNAL_DOWNLOADS]: The skill references and links to official DataDog GitHub repositories and documentation sites for templates and reference code.\n- [SAFE]: Sensitive information is protected by instructing the generated code to retrieve API and application keys from environment variables rather than hardcoding them.\n- [SAFE]: A data sanitization step is implemented to scrub PII (emails, phone numbers, SSNs, and API keys) from local JSON or CSV datasets before they are processed or embedded into the generated experiment code.
Audit Metadata