Prepare the PDF
Text-based PDFs work best. Scans, complex columns, repeated headers, and encrypted files may need OCR or cleanup first.
- Confirm that you have the right to process it
- Use a PDF without encryption
- Preserve headings and paragraph structure
- Keep the file below 50 MB
Review three things before download
Check that names do not overlap, descriptions explain the trigger, and the operating steps preserve important source constraints.
- Coverage and number of skills
- Triggers and steps in key skills
- Source title and output language
- A trial in an isolated Codex environment
Install with a rollback path
The Book2Skill installer backs up an existing skill with the same name and then replaces skills atomically.
- Download the ZIP
- Read the README
- Run deploy-mac.sh
- Inspect the backup and result
- Test one real task after installation
Common questions
Can I convert scanned PDFs?
Only after reliable OCR. Recognition quality directly affects source fidelity.
Can I edit the result?
Yes. The package contains ordinary folders and Markdown files that you can review and change before installation.
Ready to turn a PDF into Codex Skills?
Start with a file you know well, inspect the result, and decide whether it belongs in your workflow.
Upload a PDF