Git workflow

Skill pablodiazjorge/prompt-forge/packages/custom/.github/skills/git-workflow

Drop-in toolkit that gives AI coding agents persistent memory across sessions. Agent Skills, cross-session learning loop, and knowledge persistence. Zero dependencies, no backend.

Install

npx -y skills add pablodiazjorge/prompt-forge --skill git-workflow

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Git best practices for AI coding agents. Use when committing changes, creating branches, splitting PRs, reviewing code, or setting up worktrees. Covers Conventional Commits v1.0.0 spec, atomic commits, stacked PRs, squash-based workflow, and worktree patterns.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

2.5 KB, 572 tokens by cl100k_base, as published. Nobody here has run it

Git Workflow

Conventional Commits v1.0.0

<type>[optional scope][!]: <description>

[optional body]

[optional footer(s)]

Types

Type	SemVer	Use
`feat`	MINOR	New feature
`fix`	PATCH	Bug fix
`docs`	—	Documentation only
`style`	—	Formatting, no logic change
`refactor`	—	Code change, no fix/feature
`perf`	—	Performance improvement
`test`	—	Adding/correcting tests
`chore`	—	Build, deps, tooling
`ci`	—	CI/CD config
`revert`	—	Reverting commits

BREAKING CHANGE (MAJOR)

Two valid forms per spec:

feat(api)!: send email on shipment       # ! before :

feat: allow config extension

BREAKING CHANGE: `extends` key now used for extending configs

Revert

revert: let us never speak of the noodle incident

Refs: 676104e, a215868

Atomic Commits

One logical change per commit. Use git add -p for interactive staging:

git add -p                     # Stage hunks interactively
git add src/models/            # Stage specific directory

Good: "Add user model and validation" / Bad: "Add user model, fix navbar bug, update readme"

PR Splitting (Stacked PRs)

PR 1 (base): Models, interfaces, types
PR 2 (logic): Services, state management
PR 3 (UI): Components, templates, styles
PR 4 (tests): Unit/integration tests

Each PR compiles independently. Squash-based workflow: lead maintainer cleans commit messages on merge.

Branch Naming

feature/<name>   fix/<name>   refactor/<name>   docs/<name>   chore/<name>

Worktrees (for PR reviews)

git worktree add -b review-pr-42 ../review feature/branch
git worktree list
git worktree remove ../review

Stashing

git stash                              # Quick
git stash push -m "WIP: refactoring"   # Named
git stash --include-untracked          # + untracked
git stash --staged                     # Only staged
git stash pop                          # Apply latest

Gives 1 of the 12 instructions most pr commit review skills give in 572 tokens

Counted across 888 of the 1,342 authors here whose files we hold, read 2026-08-07

use conventional commits formatin 127 of 888, across 115 files
keep subject line under 72 charactersin 62 of 888, across 48 files
delete branches after mergein 51 of 888, across 38 files
use imperative mood in subject linein 51 of 888, across 42 files
use imperative mood in commit messagesin 44 of 888
verify directory is ignored before creating worktreein 43 of 888, across 12 files
generate a conventional commit messagein 43 of 888
add unignored worktree directories to gitignorein 42 of 888, across 10 files
make atomic commitshere, and in 39 of 888, across 27 files
run tests before committingin 36 of 888, across 25 files
verify clean test baselinein 35 of 888, across 9 files
split unrelated changes into separate commitsin 35 of 888, across 30 files

Said here and by no other author read

stage hunks interactively
ensure each pull request compiles independently
use git worktrees for pull request reviews

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.