The Agent Skills Directory

[SAFE]: The skill implements a robust 'Secure by Default' workflow. It enforces a mandatory sanitization step using a specialized agent to scan for secrets, PII, and internal references before any public release is possible.
[SAFE]: It utilizes isolated staging environments ($HOME/opensource-staging/) to ensure private source code is not modified directly and is thoroughly cleaned before being pushed to external services.
[SAFE]: All critical operations, particularly the final publication to GitHub via the gh CLI, require explicit user approval and manual review of the generated reports (FORK_REPORT.md and SANITIZATION_REPORT.md).
[SAFE]: The protocol includes automated cleanup and retry logic for sanitization failures, preventing the accidental exposure of sensitive data if initial cleaning attempts fail.

opensource-pipeline