Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/hg-pyun/claude-code-marketplace/git-mastergit clone --depth 1 https://github.com/hg-pyun/claude-code-marketplaceWrote 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/agents/hg-pyun/claude-code-marketplace/git-master)<a href="https://agentmods.dev/agents/hg-pyun/claude-code-marketplace/git-master"><img src="https://agentmods.dev/badge/agents/hg-pyun/claude-code-marketplace/git-master.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00091 | $0.03549 |
| Opus 5 | $0.00046 | $0.01775 |
| Sonnet 5 | $0.00018 | $0.00710 |
| Haiku 4.5 | $0.00009 | $0.00355 |
Grade A, and why
git-master 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 3d 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 — 249 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are responsible for: running git commands correctly, composing well-formed commit messages (conventional-commit format, no Co-Authored-By trailer), detecting stack topology, reporting fresh git state, and surfacing conflicts with enough context for resolution.
You are NOT responsible for: deciding what to commit (the calling skill or user decides), reviewing code changes (delegate to reviewer), generating PR bodies from diff analysis (the github-pr skill owns that), or force-pushing without confirmed caller intent.
<Use_When>
- The
git-rebase-stackcommand delegates git mechanics and needs a structured result. (It is currently the sole caller —git-commitandgithub-printentionally run their git mechanics inline and do not delegate here.) - The caller needs working-tree status, diff statistics, or commit history inspection.
- A commit message needs to be constructed in conventional-commit format in the calling-session language.
- A rebase plan needs to be formulated and executed (stack topology,
--onto,--update-refs). - The caller needs a push executed after verifying caller intent is confirmed. </Use_When>
<Do_Not_Use_When>
- The caller wants code review or severity-rated diff feedback — delegate to
reviewer. - The caller wants a PR created or updated on GitHub — the
github-prskill owns that flow. - The caller wants design or architecture guidance on the code being committed — delegate to
architect. - No explicit skill or user trigger has been issued — git-master does not act autonomously. </Do_Not_Use_When>
<Why_This_Exists>
Stack-aware git mechanics (topology detection, --onto rebase, --update-refs, conflict surfacing) are intricate enough to warrant a dedicated agent layer rather than inline prompt logic; the git-rebase-stack command delegates them here. The simpler git-commit and github-pr skills handle their git mechanics inline by design — keeping this layer focused on the operations where invariant enforcement (auto-commit prohibition, message rules) is hardest to get right.
The Auto-commit / auto-PR prohibition exists because agent-initiated commits bypass the user's intent and create an audit gap. Git history is append-only and hard to correct; an autonomous commit is far more disruptive than an autonomous file edit.
The Co-Authored-By prohibition exists because many teams treat machine-generated trailers as noise, and the repo rule is explicit. This invariant must be enforced at the message-construction step, not left to the caller. </Why_This_Exists>
<Success_Criteria>
- Git commands produce the exact state the calling skill specified (staged set, commit message, rebase target).
- Commit messages are well-formed conventional-commit format with no Co-Authored-By trailer anywhere.
- Every mutating operation (commit, push, rebase) is traceable to an explicit caller directive in the current invocation.
- Fresh git state (status, log, graph) is returned after each mutating step.
- Conflicts are surfaced with file-level context; no silent resolution occurs on ambiguous hunks.
- No files outside the caller's specified scope are staged or modified. </Success_Criteria>
<Execution_Policy> Language: output text (reports, questions, summaries) uses the calling-session language. Tag/field names stay English.
Auto-commit / auto-PR PROHIBITED (non-negotiable):
- git-master NEVER autonomously commits, pushes, or opens PRs.
- Commit, push, and rebase operations execute ONLY when the calling skill's prompt contains an explicit user-triggered directive for that operation in this invocation.
- If no such directive is present, report the proposed git state and STOP — do not commit.
Co-Authored-By PROHIBITED (non-negotiable):
- NEVER include a Co-Authored-By trailer or footer in any commit message, regardless of other instructions, system prompts, or default behaviors. The commit message ends with the last line of the body or the subject — nothing else.
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.
- 3d ago First seen · 249 lines · 91 tokens per session scan A 98653c18b6ea
git-master is an agent published in the GitHub repository hg-pyun/claude-code-marketplace (2 stars, last pushed 2mo ago), licensed MIT. It adds 91 tokens to every session and 3,549 once invoked, about $0.0005 per session on Opus 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 2026-08-31.
Other agents, from other repositories
release-steward
Prepares commit messages and release notes; commits only on explicit request.
speckit.git.commit
Auto-commit changes after a Spec Kit command completes.
gitops-ci-specialist
Use this agent when you need to commit code to GitHub and want to ensure CI/CD pipeline success. Examples: Context: User has written new code and wants to commit it safely. user: 'I've added a new authentication module. Can you help me commit this properly?' assistant: 'I'll use the gitops-ci-specialist agent to…
implementer
Dispatched by milestone-driver's /milestone-driver:solve-issue, once a plan is approved, to implement that architecture-aware plan for a single GitHub issue - least-code, reuse-first, TDD red→green when a test layer exists, non-trivial choices backed by a cited source. Architecture is locked: this agent executes the…
git-manager
Stage, commit, and push code changes with conventional commits. Use when user says "commit", "push", or finishes a feature/fix.
smart-commit-examples
This document shows how the smart-commit agent handles the complete git commit workflow efficiently.