git
Load when planning or about to git commit, branch, push, rebase, resolve a merge conflict, open PR, stack PRs (even when gh-stack drives the stack), or…
Skill by sherifabdlnaby
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/sherifabdlnaby-git
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: devops, writing
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/git
What's inside
- Skill name
git- Description
- Load when planning or about to git commit, branch, push, rebase, resolve a merge conflict, open PR, stack PRs (even when gh-stack drives the stack), or write/update the description of a Pull Request (PR), attach a screenshot or video to a PR, respond to review comments, review a PR, watch/babysit a PR's CI or reviews. Carries the user's git conventions and style you are expected to match. Load as early as possible, the moment git or GitHub work is anywhere in the chat's future.
- License
- MIT
- Metadata
- author: sherifabdlnaby · version: 0.5.0
- Instructions
- 142 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (14)
- .gitignore
- README.md
- SKILL.md
- attach.md · references/
- branches.md · references/
- commits.md · references/
- pr-body.md · references/
- pull-requests.md · references/
- rebase.md · references/
- review-responses.md · references/
- reviewing.md · references/
- watch-flags.md · references/
- watch.md · references/
- pr-watch.py · scripts/
SKILL.md (first lines)
# Git Route first: each action's actual rules live in a `references/` file. Open the matching one before planning or acting; rules like commit grouping and branch naming shape the plan, so reading them late means redoing it. Several actions, several references. ## Router **Branches** (create, name, stack) -> [`references/branches.md`](references/branches.md) Switching branches, naming rules, ticket references, switching with a dirty tree, stacked PRs (conventions here, mechanics via the gh-stack skill), the gh-stack/manual path probe, cascade conflicts and verification, manual restacking fallback (after merge, chain of 3+, mid-stack push), force-pushing. **Committing** -> [`references/commits.md`](references/commits.md) Pre-staging walk, file confirmation, message style, hook-failure handling, scoping, editor-free squash, empty commits. **Pull Requests** (open, update) -> [`references/pull-requests.md`](references/pull-requests.md) Pre-flight survey, title format (including the stacked `[n/N]` marker), `gh pr create` flags, post-create lines, finishing PRs after `gh stack submit`, updating a body without clobbering it. **PR body** (write or rewrite a description) -> [`references/pr-body.md`](references/pr-body.md) Voice for the product and engineer blocks, each block's rule in body order, where a visual sits and when to take a screenshot, receipts and the structure hook, the skeleton, linking, AI footers. **Attach a screenshot or video** -> [`references/attach.md`](references/attach.md) `--attach` on create/edit/comment, body-file placement, screenshot framing guidance. **Responding to review comments** (on your own/target PR) -> [`references/review-responses.md`](references/review-responses.md) Classify the reviewer (automated bot/AI tool, an AI-disclosed agent behind a human account, or a human), then fix / push back / escalate; in-thread replies, thread resolution, re-requesting review, batching a round into one GraphQL query and one aliased mutation, where a PR stands. **Rebasing, merging, pulling, resolving conflicts** -> [`references/rebase.md`](references/rebase.md) Rebase vs merge, how to pull, who resolves which hunk, conflict guidance, force-with-lease, verification. Stacked-PR restack live in [`references/branches.md`](references/branches.md); read it too when the branch is in a stack. **Reviewing someone else's PR** -> [`references/reviewing.md`](references/reviewing.md) Procedure: gather context, check out locally, trace the change, deliver a summary and review guide. **Watch a PR's CI and reviews** -> [`references/watch.md`](references/watch.md) The four verdicts, the three modes (green CI / green and quiet / reviews only), the cheap sub-agent that runs the loop, the stale nudge. Uses `scripts/pr-watch.py`, never a sleep loop; its flags are in `references/watch-flags.md`. **Stuck CI check with no manual retry** -> [`references/commits.md`, Empty commits](references/commits.md#empty-commits) The last-resort empty commit: when it beats retrying the check, and the zero-file-change guards that keep approvals alive. ## Always **Disclose AI.** Anything posted on GitHub on the user's behalf (PR body, comment, issue, ticket update) carries an **AI footer**. Commit messages carry no footer. The PR-body template lives in [`references/pr-body.md`](references/pr-body.md); the post templates, one per tier of human judgment behind the post (Agent Decided, Human Planned, Human Guided), live in [AI Disclosure](#ai-disclosure) below. Use them as verbatim as possible, do not write from memory. **Reach for `gh` first.** Every GitHub action goes through the `gh` CLI; the GitHub MCP server is the fallback for what `gh` cannot do. **Parallelize read-only calls.** Batch independent read-only `git`/`gh` calls (`status`, `diff`, `log`, `gh pr view/diff/checks`) into a single tool call; each sequential read is a roundtrip for nothing. Mutating commands (`commit`, `push`, `rebase`, `gh pr create/edit/merge`) stay sequential. **Fewer round-trips, smaller payloads.** Batching is the first lever; these are the next four, your call when a one-off is cheaper: - **One wide fetch, many narrow reads.** Anything you'll consult more than once in a task (PR JSON, `gh stack view --json`, a review-thread query) goes to a temp file once, and each step reads its slice with `jq`. Re-fetch after anything you did that changes it (push, edit, reply, resolve); or if time passed and you've been idle, a snapshot is not a live view. - **Ask for an answer, not a payload.** A count or a single field leaves as `--jq '.x'`; a yes/no leaves as an exit code (`git diff --quiet`, `git merge-base --is-ancestor`). JSON you read with your own eyes is a payload. - **Same-shape mutations ride one request.** GraphQL aliases: every reply, resolve, or title edit of a round in one `gh api graphql`. Aliases fail independently, so read the per-alias `errors[].path`, not the exit code. - **Bodies travel by file.** PR body or GraphQL query to a file, then `--body-file` / `-F query=@file`. (why: shell quoting eats backticks and `$`, and a file stays editable for the next update.) Hooks that ship beside this skill check PR bodies deterministically, and they refuse a body they cannot read from a file. **Resolve once per session, reuse everywhere.** One line for all three: ```bash gh repo view --json nameWithOwner,defaultBranchRef --jq '"repo \(.nameWithOwner)\ntrunk \(.defaultBranchRef.name)"'; gh api user --jq '"user \(.login)"' ``` Carry `<OWNER/REPO>`, `<trunk>` and `<GITHUB_USERNAME>` through the session from that. **Voice** for everything public (commit messages, PR titles and bodies, comments, issues): - Omit needless words. Concise sentences, no padding paragraphs. - Emojis sparingly, only where one helps the reader catch something while glancing. - NO em dashes in any public-facing text. Use commas, parentheses, or periods. (why: em dashes are a classic AI tell and repel readers.) - No vague reaffirmations like "for accountability", "for performance", "for resiliency" unless that reason is already in the conversation. (why: invented justification misrepresents the change.) - Dry, low-key humor. - Don't overdo formatting. Keep it balanced. (why: visual noise crowds out the content.) - A user-supplied Human Note is exempt: it goes in the PR body verbatim. ## AI Disclosure Every post on the user's behalf (comment, reply, issue) ends with an AI footer: a `---`, then one line copied verbatim from the templates below. The PR-**body** variants live in [`references/pr-body.md`](references/pr-body.md) and follow the same rule. Placeholders, the same in every footer: - `<TOOL>`: the tool you're running as, e.g. Claude, Cursor. - `<MODEL>`: the friendly name of the model you're running, e.g. `Opus 4.8`. - `<GITHUB_USERNAME>`: resolve once per session with `gh api user --jq '.login'`. ### Picking the variant The footer answers one question for whoever reads the post: how much human judgment stands behind it? Judge the decisions the post carries, not the request that produced them. The tiers go from least to most human judgment. **🤖 Agent Decided.** You chose the change, position, or wording. No human guidance on the how of it. ```markdown _<sub>🤖 Agent Decided: Posted by <TOOL> (<MODEL>) autonomously on behalf of @<GITHUB_USERNAME>.</sub>_ ``` **📋 Human Planned.** The user worked through the plan with you, in a grill session or a plan that covers many decisions, and the post carries it out. The implementation details are yours. A short "go do this" is not a plan: that stays 🤖. ```markdown _<sub>📋 Human Planned: Posted by <TOOL> (<MODEL>) on behalf of @<GITHUB_USERNAME>.</sub>_ ``` **🤝 Human Guided.** A back and forth on a technical decision happened, and its outcome is in this post: the user picked one of the options you laid out, answered the question that settled the
From sherifabdlnaby/skills, MIT.
More from sherifabdlnaby/skills
- 0.633score
mise-fy
sherifabdlnaby
Use when working with mise (mise-en-place) or planning to update mise.toml.
SkillAFreshness: A★ 9Claude Code +9 - 0.633score
cicd-fy
sherifabdlnaby
Use when setting up, shaping, or auditing a project's CI/CD: a build/test/scan pipeline, a release pipeline, or publishing an artifact (container image…
SkillAFreshness: A★ 9Claude Code +9 - 0.629score
skill-fy
sherifabdlnaby
Load when creating a new skill, scaffolding a SKILL.md, adding a reference page to a skill, writing or tuning a skill (or their description) or…
SkillAFreshness: A★ 9Claude Code +9 - 0.622score
coding
sherifabdlnaby
The user's coding and docs conventions and taste.
SkillAFreshness: A★ 9Claude Code +9 - 0.619score
afk
sherifabdlnaby
Act autonomously while I'm away.
SkillAFreshness: A★ 9Claude Code +9 - 0.619score
fast-feedback-mode
sherifabdlnaby
Act fast, keep me in the zone.
SkillAFreshness: A★ 9Claude Code +9 - 0.619score
kubernetes-debug
sherifabdlnaby
Use for Kubernetes Debugging Sessions
SkillAFreshness: A★ 9Claude Code +9 - 0.619score
review
sherifabdlnaby
Gate the work behind a local reviewer panel.
SkillAFreshness: A★ 9Claude Code +9 - 0.619score
watch-pr
sherifabdlnaby
Watch a PR until CI is green and its reviews are answered.
SkillAFreshness: A★ 9Claude Code +9 - 0.617score
watch-pr-comments
sherifabdlnaby
Watch a PR's reviews and comments and answer them as they land.
SkillAFreshness: A★ 9Claude Code +9 - 0.614score
watch-pr-ci
sherifabdlnaby
Watch a PR's CI until it is green, fixing what goes red.
SkillAFreshness: A★ 9Claude Code +9 - 0.614score
afk-yolo
sherifabdlnaby
/afk, max non-destructive autonomy.
SkillAFreshness: A★ 9Claude Code +9
Signals
- Stars
- 9
- Forks
- 1
- Open issues
- 0
- Last commit (this folder)
- Sep 25, 2026 (1 month ago)
- Last release
- Oct 23, 2026 (5 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- Version
- 0.5.0
- License
- MIT
- Repository
- github.com/sherifabdlnaby/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from sherifabdlnaby
- 0.633score
cicd-fy
sherifabdlnaby
Use when setting up, shaping, or auditing a project's CI/CD: a build/test/scan pipeline, a release pipeline, or publishing an artifact (container image…
SkillAFreshness: A★ 9Claude Code +9 - 0.633score
mise-fy
sherifabdlnaby
Use when working with mise (mise-en-place) or planning to update mise.toml.
SkillAFreshness: A★ 9Claude Code +9 - 0.629score
skill-fy
sherifabdlnaby
Load when creating a new skill, scaffolding a SKILL.md, adding a reference page to a skill, writing or tuning a skill (or their description) or…
SkillAFreshness: A★ 9Claude Code +9 - 0.622score
coding
sherifabdlnaby
The user's coding and docs conventions and taste.
SkillAFreshness: A★ 9Claude Code +9 - 0.619score
fast-feedback-mode
sherifabdlnaby
Act fast, keep me in the zone.
SkillAFreshness: A★ 9Claude Code +9 - 0.619score
review
sherifabdlnaby
Gate the work behind a local reviewer panel.
SkillAFreshness: A★ 9Claude Code +9
Cite this page
git skill. Agent Mart. https://agentmart.dev/skills/sherifabdlnaby-git (last updated 2026-10-24)