Install
$ agentstack add skill-pfeff-claude-skills-git ✓ 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 Skill
Encapsulates preferred git usage patterns and conventions.
Core Principles
Incremental Commits: Commit at logical boundaries during implementation
| Commit when... | Don't commit when... | |----------------|---------------------| | Logical unit complete (model, service, component) | Small part of a larger unit | | Tests pass + meaningful progress | Tests failing | | About to switch contexts (backend → frontend) | Purely scaffolding with no behavior | | About to attempt risky/uncertain changes | Would need a "WIP" commit message |
Heuristic: "Can I write a meaningful commit message? If yes, commit. If I'd write 'WIP', keep working."
Commit Messages: Keep compact and meaningful
- Headline + 1-2 sentence summary maximum
- Avoid duplicating what's obvious from the diff
- For bug fixes: indicate the specific bug that was fixed
- Use conventional commit format (feat, fix, refactor, docs, etc.)
- NO "Generated with Claude" or "Co-Authored-By" annotations
Operations
Committing Changes
When: User requests a commit or asks to save work to git.
Implementation: Load operations/commit.md for detailed steps.
Quick summary: Stage changes, craft compact commit message, create commit.
Merging Pull Requests
When: User requests merging a PR, especially from a worktree context.
Implementation: Load operations/merge-pr.md for detailed steps.
Quick summary: Resolve merge strategy (explicit input → repo CLAUDE.md Merge Strategy → GitHub repo settings → default --merge), detect worktree context, merge without --delete-branch in worktrees to avoid checkout failure, delete remote branch separately.
Submodule Workflows
When: Working with git submodules, especially after clone or dotfiles install.
Implementation: Load operations/submodule.md for detailed steps.
Quick summary: Check for detached HEAD, sync to main before making changes, commit submodule then parent.
Creating Pull Requests
When: User requests a PR for the current branch.
Implementation: Detect PR tool based on remote origin, then use appropriate command.
Detection:
# Detect whether to use gh or az for PRs
tool=$(${CLAUDE_PLUGIN_ROOT}/skills/git/bin/detect-pr-tool)
PR Creation by Tool: | Tool | Command | |------|---------| | gh | gh pr create --title "..." --body "..." | | az | az repos pr create --title "..." --description "..." |
Quick summary: Run detect-pr-tool to determine gh vs az, then create PR with appropriate CLI.
Commit Message Format
:
Types:
feat: New featurefix: Bug fixrefactor: Code restructuring without behavior changedocs: Documentation onlytest: Test additions or modificationschore: Maintenance tasks
Examples:
Good:
fix: correct AD username rendering in init script
The username variable was incorrectly escaped, causing registration failures.
feat: add OAuth2 support for API authentication
Enables token-based auth for external integrations.
Bad (too verbose):
fix: fixed the bug where the AD username was being rendered incorrectly
This commit fixes the issue where the AD username variable in the init script
was being rendered with incorrect escaping, which was causing the servers to
fail to register with Active Directory and subsequently fail all downstream
initialization steps.
Progressive Disclosure
Load only what you need:
Operations:
operations/commit.md- Commit workflow implementationoperations/merge-pr.md- Worktree-aware PR mergeoperations/submodule.md- Submodule workflow patterns
Scripts:
bin/detect-pr-tool- Detectghorazbased on remote origin
Templates:
templates/commit-message.tmpl- Message structure template
See Also
- User's global CLAUDE.md for commit conventions
- Conventional Commits specification
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: pfeff
- Source: pfeff/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.