Skip to content
Back to skills

Commit

ASecurity

Create focused conventional commits with a mandatory gitmoji. Use when the user asks to commit, write a commit message, or split changes into commits.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 22, 2026
ai-agentsrailsgit

Works with

  • cli

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned October 22, 2026

npx -y skills add skrrt-sh/skills --skill commit --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Commit?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Commit
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/skrrt-sh-commit/badge)](https://www.skillsdirectory.com/skills/skrrt-sh-commit)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: commit
description: Create focused conventional commits with a mandatory gitmoji. Use when the user asks to commit, write a commit message, or split changes into commits.
---

# Commit Changes

Create the smallest coherent commits that explain what changed and why.

## Format

```text
type(scope): :gitmoji: imperative subject

body explaining what changed and why

footer
```

- Use a standard conventional-commit type and [gitmoji](https://gitmoji.dev) code.
- Put the gitmoji immediately after `type(scope):`; omit an uncertain scope.
- Include a body for every commit.
- Add `BREAKING CHANGE:` only for a real breaking change.
- Add issue footers only for issues the user or repository identifies.
- End with `Co-Authored-By: Skrrt Bot <bot@skrrt.sh>` unless the user opts out.

## Process

1. Read `.agents/ship.md` at the repository root. If it is missing in a repository that is not
   the user's own, such as open source or an org-wide repo, never run setup: follow its
   contributing rules and treat its default branch as protected. If the user declined setup, do
   the same without asking again and report that no policy is configured. Otherwise stop and ask
   the user to run `/setup`.
2. Inspect `git status --short`, staged and unstaged diffs, and the current branch. Do not include
   unrelated files without approval.
3. Protect work on a forbidden branch before staging: create a correctly named work branch from
   the current `HEAD` with `git switch -c <type>/<description>`. Do not pull, switch bases, or stash
   first; the new branch must retain the current worktree exactly.
4. If this is a fix to a recent PR or MR, read
   [references/follow-up.md](references/follow-up.md) before staging.
5. Group the diff by intent. If a type and gitmoji disagree, split the change instead of forcing a
   message.
6. Stage only the intended files or hunks. Review `git diff --cached` before each commit.
7. Write the message to a `mktemp` file outside the repository, commit with
   `git commit --file <file>`, then remove it. A fixed path lets another commit overwrite it.
8. Verify the resulting commit and report its hash, subject, included files, and known test status.

## Guardrails

- Commit only when the user asked for a commit; loading this skill is not that request.
- Follow the protected branches and work base in `.agents/ship.md`.
- Preserve unrelated and untracked work.
- Never invent test results, amend, rewrite history, or force-push unless explicitly requested.
- Never commit secrets or likely credential files; stop and warn the user.
- Never commit absolute local paths or machine-specific config.
- If commit signing fails, keep working and retry every 30 seconds, at most five attempts in total. Then stop and
  report it in one line. Never disable signing or skip hooks.

Files in this skill

  • SKILL.md2.8 KB
  • agents/openai.yaml237 B
  • evals/evals.json5.7 KB
  • evals/trigger-evals.json1.6 KB
  • references/follow-up.md627 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.