commit-session

To commit only changes from a specific session, use a session-scoped baseline to compute and stage just the dirtied files, preserving preexisting work. This approach isolates files touched in the active session and pushes them with a conventional-commit message.

A session-scoped git commit captures only files modified within the current session by using a unique session identifier to track the baseline. It avoids accidental shipping of preexisting or unrelated dirty work by staging only the diff generated during that specific session.

Yes, you need the CODEX_THREAD_ID environment variable to identify the active session. This identifier establishes the session-scoped baseline required to accurately compute the committable changes and push them to the repository.

The best way to commit session changes without mixing preexisting work is to apply a baseline-aware diff that isolates session files. This method runs lightweight checks before staging and creates a structured conventional-commit message for the pushed changes.

Yes, session-scoped committing works across any git repository because it relies on the session identifier and a computed diff rather than repository-specific configurations. It automatically stages and pushes only the files touched in that active session.

Your git commit includes unmodified files because standard staging captures all dirty work. To prevent this, use a session-scoped baseline that specifically computes and isolates the committable changes dirtied only during the active session.