commit

Workflow

1. Run these commands in parallel to understand the current state:

   • git status — see all tracked/untracked files and staging state
   • git diff — see unstaged changes

2. Analyze ALL changes (staged + unstaged) and determine what should be committed:

   • If there are no changes at all, inform the user and stop
   • Do NOT commit files that likely contain secrets (.env, credentials, tokens, API keys). Warn the user if such files are present.

3. Group related file changes into logical commits:

   • Do NOT lump all changes into a single commit
   • Examine each changed file and understand what it does
   • Group files that are part of the same logical change together (e.g., a component + its styles + its test = one commit)
   • Each group becomes its own commit
   • If only one logical change exists, one commit is fine

4. For each group, draft a commit message:

   • Write natural, concise descriptions of what changed and why
   • Do NOT use conventional commits format (no feat:, fix:, chore: prefixes)
   • Do NOT add Co-Authored-By or any co-author trailers to commit messages
   • Keep the subject line under 72 characters
   • If the user provided a message or instructions via arguments, incorporate or use that for the relevant commit(s)

5. For each group, stage and commit:

   • git add <specific files> for that group only
   • git commit -m "message"
   • Repeat for each group in a logical order (foundational changes first)

6. Run git status after all commits to verify everything is clean.

7. Report the result: show each commit hash and summary.

User Arguments

$ARGUMENTS