commit

name	commit
description	Create conventional commits of changes. Use when asked to commit, save changes, create a commit, or when finishing implementation work. Also use when the user says "commit this", "save my work", "create commits", or just "/commit".
argument-hint	Describe the change you want to commit
model	sonnet

Create conventional commits grouped by logical changes. Scope: $ARGUMENTS (if empty, commit all changes).

Step 1: Gather context

Run these in parallel:

git status
git diff --stat
git diff --staged --stat
git log --oneline -10

This tells you what's changed, what's already staged, and what the recent commit style looks like.

Step 2: Normalize the staging area

If files are already staged from a previous partial workflow, reset them so you can regroup logically:

git reset HEAD

This is safe — it only unstages, never discards work. Skip this if nothing is staged.

Step 3: Check for secrets and forbidden files

Before staging anything, verify no secrets or sensitive files are in the changeset:

git diff --name-only | grep -iE '(secret|\.env$|credentials|token|password|key\.pem|\.key$)' || true

False positives to ignore: *.env.example, .gitkeep, secrets.tf (OpenTofu config, not actual secrets), secrets/ dirs containing only .gitkeep. Use judgment — the grep catches filenames, not file contents.

If any real secrets are found, do NOT commit those files. Warn the user and skip them.

Step 4: Review diffs and plan commit groups

For small changesets (<15 files), read the full diff:

git diff

For large changesets, use --stat output from Step 1 to understand the shape, then spot-check key files if grouping isn't obvious from paths alone.

Plan all commit groups before executing any. Write out the groups mentally — this prevents orphan files that don't fit any later commit. Each group should be a cohesive, self-contained change.

Grouping heuristics (in priority order):

1. By host — all compose files for one host = one commit (compose/mac-mini/, compose/nas/, compose/rpi4/)
2. By stack/module — a single OpenTofu stack or module = one commit, unless changes span many stacks (then group all terraform together)
3. By concern — CI workflows, Grafana dashboards, operational scripts, etc.
4. By config layer — .gitignore + .env.example together; CLAUDE.md + skills + rules together
5. Unrelated single files — separate commits or group with the nearest logical neighbor

When in doubt, fewer larger commits beat many tiny ones — a commit should tell a coherent story.

Step 5: Execute commits

For each planned group, stage and commit in one step:

git add <file1> <file2> ... && git commit -m "$(cat <<'EOF'
<type>(<scope>): <description>

Co-Authored-By: Claude <[e-mail]>
EOF
)"

Commit message format

Follow Conventional Commits:

<type>(<scope>): <short description>

Types: feat, fix, chore, refactor, docs, ci, test

Scopes (derive from the files changed):

• Host names: mac-mini, rpi4, nas
• Infra areas: tofu, ci, caddy, komodo, renovate
• Specific stacks: immich, observability, autopirate, etc.
• Use claude for CLAUDE.md/skills changes, infra for cross-cutting infra changes

Rules:

• Lowercase description, no period at the end
• Focus on why, not what — the diff shows what changed
• Keep the first line under 72 characters
• Add a blank line + body only if the why isn't obvious from the subject
• Always include the Co-Authored-By trailer

Step 6: Verify

git status
git log --oneline -5

Confirm all intended changes are committed, nothing unexpected was staged, and the log looks clean.

Rules

• NEVER push — only create local commits
• NEVER commit .env files (only .env.example is allowed)
• NEVER commit files matching *secret*, *-secret*, *secret*.env
• NEVER use git add -A or git add .
• NEVER amend existing commits unless explicitly asked
• If a pre-commit hook fails, fix the issue and create a NEW commit