agentleFS
nimadorostkarnimadorostkar/Claude-Skills-collection/skills/development/git-workflow/SKILL.md

git-workflow

Skill33 starsChanged 22 days agoMITView on GitHub

Use for branching, committing, history repair, and release hygiene. Covers atomic commits, rebase versus merge, bisect, reflog recovery, and undoing mistakes safely. Trigger on "undo my last commit", "i messed up a rebase", "squash these commits", "find which commit broke this". Do not use for pipeline and deploy automation — see the ci-cd skill.

  • Deletes or force-pushes
  • Commits and pushes

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 163 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Keep history readable and recoverable.

    line 14
  2. Distinctive:

    Commit messages that explain why, not what.

    line 40
  3. Very specific:

    Never git push --force to a shared branch.

    line 53
  4. Distinctive:

    Do not commit generated files, secrets, or large binaries.

    line 56
  5. Distinctive:

    If a secret is committed, rotating it is mandatory and rewriting history is optional.

    line 97

“Common” is measured: kept by at least 6 in 10 of the 163 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Git Workflow

Purpose

Keep history readable and recoverable. A good history is a debugging tool: it lets you bisect, revert, and explain. A bad one is a wall of "fix", "fix again", "wip".

When to Use

  • Structuring commits before opening a pull request.
  • Recovering from a bad rebase, reset, or force-push.
  • Bisecting to find the commit that introduced a regression.
  • Cleaning up a branch's history.
  • Deciding a team branching strategy.

Capabilities

  • Atomic commit construction, including interactive staging.
  • History rewriting: interactive rebase, squash, fixup, autosquash.
  • Recovery: reflog, ORIG_HEAD, git restore, cherry-pick.
  • Bisection, including automated bisection with a test script.
  • Conflict resolution and rerere.

Inputs

  • The repository state and the branch's relationship to its base.
  • Whether the branch has been pushed and whether others have based work on it.

Outputs

  • A branch whose commits each build, pass tests, and do one thing.
  • Commit messages that explain why, not what.
  • A rebase or merge that preserves the ability to revert cleanly.

Workflow

  1. Commit atomically — One logical change per commit. Use git add -p to split a messy working tree.
  2. Write the message for the future — Subject in the imperative under 72 characters; body explaining why the change was needed.
  3. Clean before review — git rebase -i to squash fixups and reorder. Reviewers should read the story, not the process.
  4. Rebase private branches, merge shared ones — Rewriting history that others have pulled is how you lose their work.
  5. Recover with reflog — Almost nothing is truly lost. git reflog shows every position HEAD has held.

Best Practices

  • Never git push --force to a shared branch. Use --force-with-lease, which refuses if someone else has pushed.
  • A commit that does not build is a commit that breaks bisect. Keep every commit green.
  • git commit --fixup <sha> plus git rebase -i --autosquash makes review-response commits disappear cleanly.
  • Do not commit generated files, secrets, or large binaries. Removing them later requires rewriting history for everyone.
  • git bisect run ./test.sh finds the offending commit automatically. It is the fastest regression hunt available.
  • Enable rerere — Git will remember how you resolved a conflict and reapply it next time.

Examples

Automated bisection:

git bisect start
git bisect bad HEAD             # current commit is broken
git bisect good v2.14.0         # this release was fine

# Any script exiting 0 for good, non-zero for bad.
git bisect run pytest tests/test_checkout.py::test_total

# Git prints the first bad commit, then:
git bisect reset

Recovering a branch destroyed by a bad reset:

git reflog                       # find the sha from before the reset
# 8a3f1c2 HEAD@{3}: commit: add proration handling
git reset --hard 8a3f1c2         # branch is back

Splitting a messy working tree into reviewable commits:

git add -p                       # stage hunks belonging to the first change
git commit -m "Extract proration calculator"
git add -p                       # stage the next logical change
git commit -m "Fix double-charge on mid-cycle upgrade"

Notes

  • The reflog is local and expires (90 days by default). It will not save you from a force-push that only ever existed on someone else's machine.
  • Merge commits make git log --oneline noisy but preserve the true topology. On a team that reviews by PR, squash-merging to main and keeping the branch history in the PR is a reasonable middle ground.
  • If a secret is committed, rotating it is mandatory and rewriting history is optional. Assume it is compromised the moment it is pushed.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.