Install
$ agentstack add skill-narumiruna-skills-git-commit ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Git Commit
Inspect the actual diff before writing the message. Prefer one coherent commit per intent, then map that intent to a Conventional Commits title and only add extra sections when they carry concrete information.
Read references/conventional-commits.md when the type is ambiguous, when you need trailers or breaking-change wording, or when the user asks for exact format details.
Workflow
- Inspect the exact commit contents.
Start with git status --short. If the user asks for a commit, inspect git diff --cached first; inspect git diff too when unstaged changes might affect the commit boundary.
- Define the commit boundary.
Group files by one user-visible intent. If the diff mixes unrelated work, recommend splitting it into multiple commits instead of forcing one vague summary over several intents.
- Choose the type and scope from behavior.
Pick the lowercase Conventional Commits type that best matches the dominant behavior. Add a short noun-like scope only when it adds signal.
- Write the title in the required format.
Use [optional scope][!]: . Keep the description short, specific, and grounded in the diff.
- Add body and footers only when they add concrete value.
Use the body for why, constraints, or tradeoffs. Add footers only for real trailers or BREAKING CHANGE:.
- Keep the message aligned with the selected diff.
Do not describe unstaged changes, planned follow-up work, or unrelated cleanup that is not part of the commit.
Decision Rules
- Prefer
featfor a new capability andfixfor incorrect behavior or regressions. - Prefer
refactorwhen behavior should stay the same, anddocswhen only documentation changes. - Prefer no scope, body, or footer unless each one adds concrete signal.
- Prefer multiple commits over one vague summary when the diff contains unrelated work.
- If the user asks for validation, check both the structure and whether the wording matches the inspected diff.
- Follow repo git rules for staging and commit structure; this skill focuses on mapping the diff to the right message.
- Do not add, preserve, or generate any AI attribution trailers in commits, PRs, or messages, including
Co-authored-by,Generated-by,Created-by,Coding-Agent, model names, tool names, or agent attribution footers.
Reference
references/conventional-commits.md: exact format, common types, footer rules, breaking-change wording, and examples.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: narumiruna
- Source: narumiruna/skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.