sync-skills-public

To sanitize Claude Code skills before publishing, you can run an export workflow that scans for sensitive content, applies sanitization rules to replace personal data, verifies the output is clean, and syncs the files to a destination git repository.

Yes, you can preview sanitized skill exports without writing files by running the export command with a dry-run flag, which simulates the scanning and sanitization process and reports changes without pushing to the destination repository.

The best way to scan local skill configurations for sensitive content is using an automated sensitive-content-scanner that checks files for private paths and personal data before applying configured sanitization rules for public sharing.

Yes, you need a local configuration file at ~/.claude/sync-skills-config.md to define sanitization rules and destination repository details before syncing sanitized skills to a public git repository.

Yes, you can scan for sensitive content in skills without a full export by running the export command with a scan-only flag, which checks files for private data and reports findings without applying sanitization or syncing to a repository.

If sensitive content is detected during the skill export process, the workflow applies configured sanitization rules to replace personal data and private paths, then verifies the output is clean before syncing to the destination repository.