Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/Kevin-Liu-01/Agent-Machinesnpx agentmods add skills/kevin-liu-01/agent-machines/git-workflowWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/kevin-liu-01/agent-machines/git-workflow)<a href="https://agentmods.dev/skills/kevin-liu-01/agent-machines/git-workflow"><img src="https://agentmods.dev/badge/skills/kevin-liu-01/agent-machines/git-workflow/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/kevin-liu-01/agent-machines/git-workflow"><img src="https://agentmods.dev/badge/skills/kevin-liu-01/agent-machines/git-workflow.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00045 | $0.00646 |
| Opus 5.5 | $0.00018 | $0.00258 |
| Sonnet 5.5 | $0.00009 | $0.00129 |
| Haiku 4.5 | $0.00005 | $0.00065 |
Grade A, and why
git-workflow scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Git Workflow
Switch over checkout
Use git switch for branch operations. Never git checkout. checkout -- silently destroys uncommitted work, and checkout <branch> is ambiguous when a file shares the name. switch only switches branches. Use git restore for file-level operations.
Trunk-based, small PRs
Target ≤ 200 lines of changed code per PR (excluding generated files). Larger changes get split into stacked diffs. Stack with worktrees, not branch-switching:
git fetch origin && git switch dev && git pull --ff-only
mkdir -p ../wt
git worktree add -b feat/a ../wt/a dev
git worktree add -b feat/b ../wt/b feat/a # stacked on feat/a
After feat/a merges, rebase downstream branches:
git -C ../wt/b fetch origin && git -C ../wt/b rebase origin/dev
Conventional Commits
type(scope): description. Type is feat, fix, refactor, docs, test, chore. Imperative mood, lowercase, no period, ≤72 chars.
feat(auth): add org-level permissions
fix(billing): handle zero-balance edge case
refactor(api): extract validation to middleware
Never use "and" in a commit message. If you need "and", you're describing two changes that should be two commits. Split them.
Force pushes
After a rebase, use --force-with-lease, never --force. Never force-push to main/master.
Heredoc for multi-line commit messages
git commit -m "$(cat <<'EOF'
type(scope): one-liner subject
Optional body explaining why, not what.
Refs #123
EOF
)"
Pre-commit checklist
git status -sb-- review staged + unstaged changes.git diff-- eyeball every hunk.git add -p-- stage precisely; split unrelated hunks across commits.pnpm typecheck && pnpm lint && pnpm test(or the project's equivalent).git commit -m "..."-- Conventional, ≤72 chars.
Never
--no-verify,--no-gpg-sign(unless explicitly requested).git rebase -ifrom a tool call (interactive, will hang).--amendonce pushed (unless explicitly requested + force-with-lease).git configupdates from a tool call.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 78 lines · 45 tokens per session scan A c18716006be6
git-workflow is a skill published in the GitHub repository Kevin-Liu-01/Agent-Machines (30 stars, last pushed 3d ago), licensed MIT. It adds 45 tokens to every session and 646 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
papi-build
Invoke when executing a build via buildexecute. Covers branching, gestalt pre-build check, and post-build audit.
papi-auto
Continue an approved PAPI cycle through implementation, verification and review using bounded autonomy. Resume durable work without rebuilding it; propose new scope before persisting it.
mcore-split-pr
Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.
engineer-bot-ops
A workflow for carrying an engineering task from scope through implementation and reviewable delivery. It works with a repository's branch, existing changes, rules, checks, and permissions while separating completed, unverified, and next-step work.
effect-ledger
Record an effect, sign an effect, reconcile an effect or check "did this PR merge" through the signed AIWG effect ledger, using one aiwg effect command with scriptable exit codes.
omh-ralph-task
Execute one omh-ralph task: file-scope, commit, report.