cookbook-audit
Pass
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local utility script,
validate_notebook.py, to automate the technical audit of Jupyter notebooks. This execution is a standard functional component provided by the trusted author. - [PROMPT_INJECTION]: The skill processes untrusted content from user-provided notebooks, creating a surface for indirect prompt injection. This is a known risk inherent to data-processing tools, but is mitigated by the skill's design. 1. Ingestion points: Jupyter notebook files processed by the validation script. 2. Boundary markers: The workflow transitions from automated output to a manual review process based on a style guide. 3. Capability inventory: Executes local Python scripts for validation. 4. Sanitization: Includes automated scanning for credentials and technical quality checks.
Audit Metadata