Commit Unstaged
Checks unstaged changes into Git as a sequence of commits with high quality commit messages. Use this whenever the user asks you to commit changes.
Install
$ agentstack add skill-evnm-claude-skills-commit-unstaged ✓ 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.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-evnm-claude-skills-commit-unstaged) 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
Read the unstaged changes (i.e. the output of git diff, plus the contents of any untracked files), think about code change(s) they represent, and then help me check the changes in to Git in a series of commits.
This interaction occurs in two parts:
- First, based on your review of the changes, propose a series of commits.
- Each commit should be self-contained and include a high quality commit message. A good commit message is one that conveys enough information for someone casually familiar with our codebase to understand it. Without being overly verbose, include any necessary context, a description of the problem or gap the commit addresses, and the solution the commit adds to the codebase.
- When writing commit messages, adhere to the commit message guidelines in this repository's CLAUDE.md.
- Collectively, the commits should be in a coherent order. We want to make it easy for code reviewers understand the broader change represented by a series of commits. Early commits might add independent components or tools that are then built upon by later commits, culminating in a "main" commit which uses those components/tools to solve the problem addressed by a pull request.
- Model the presentation of the plan on the output of
git rebase -i. Show the ordered list of commit messages. - Always end step 1 with the AskUserQuestion tool to get my explicit confirmation before committing anything — never proceed to step 2 based on plain-text approval alone, even for a single obvious commit. Offer at minimum "Commit as proposed" and "Revise message" (or similar) as options.
- Second, once you've received by approval on the proposed commits, run the
git add,git commit, etc commands necessary to enact the series of commits.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: evnm
- Source: evnm/claude-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.