cookbook-audit

Automated notebook auditing runs technical validation against a defined rubric, checking for issues like hardcoded secrets, suppressed outputs, and code quality problems. This Skill executes python3 validate_notebook.py to scan your notebook, applies detect-secrets for credential detection, and generates a structured audit report with actionable findings.

Yes. This Skill uses detect-secrets to scan notebooks for exposed API keys and credentials before publication. It integrates secret scanning into the audit workflow, flagging security risks alongside code and narrative quality issues in a single comprehensive report.

Run the notebook through automated validation using this Skill to evaluate narrative clarity, code quality, technical accuracy, and actionability against a rubric. It converts the notebook to markdown for context-rich review and produces an executive-style audit report with specific improvement areas.

Converting notebooks to markdown via nbconvert creates a human-readable format optimized for review workflows. This Skill applies conversion as part of its audit process, making it easier to spot narrative gaps, verify technical accuracy, and provide feedback without reopening the original notebook file.

The audit evaluates four dimensions: narrative quality, code quality, technical accuracy, and actionability. The Skill scores each category against a defined rubric and generates a structured report identifying gaps and recommendations specific to each area.

Automated auditing reduces manual review time by pre-screening notebooks for rubric violations and security risks. It flags issues systematically and generates detailed reports, but human review of recommendations and edge cases remains part of a complete publication workflow.