Borrowing it
Nothing to install: this file belongs to selfagency/beans-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/selfagency/beans-mcp/main/.agents/skills/git-mcp-workflow/SKILL.mdgit clone --depth 1 https://github.com/selfagency/beans-mcpWrote 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/selfagency/beans-mcp/git-mcp-workflow)<a href="https://agentmods.dev/skills/selfagency/beans-mcp/git-mcp-workflow"><img src="https://agentmods.dev/badge/skills/selfagency/beans-mcp/git-mcp-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/selfagency/beans-mcp/git-mcp-workflow"><img src="https://agentmods.dev/badge/skills/selfagency/beans-mcp/git-mcp-workflow.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00160 | $0.02683 |
| Opus 5.5 | $0.00064 | $0.01073 |
| Sonnet 5.5 | $0.00032 | $0.00537 |
| Haiku 4.5 | $0.00016 | $0.00268 |
Grade A, and why
git-mcp-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 2mo 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 — 219 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Git MCP Workflow
An MCP-first workflow skill for repositories exposed through git-mcp.
This skill adapts common Git guidance for a world where the preferred interface is the server's tool catalog, not ad-hoc shell commands. It covers both everyday Git work and the spicier recovery and history-editing cases.
Why LLMs Must Not Use the Git CLI
Using git directly via the shell is highly error-prone for LLMs and should be considered forbidden except as a narrow last resort.
The root cause is structural: Git commands frequently require commit messages, branch names, file paths, and other string arguments that contain spaces, quotes, special characters, or newlines. Shell interpreters handle quoting through multiple layers (single quotes, double quotes, $'...' escapes, heredocs) and the rules differ between sh, bash, and zsh. LLMs routinely produce subtly malformed quoting that causes:
- Commit messages truncated at the first space or quote character
- Arguments interpreted as flags (a message starting with
-becomes a flag) - Multi-line messages collapsed to one line or silently dropped
- Shell word-splitting breaking file paths that contain spaces
- Quote nesting errors causing the command to hang waiting for a closing delimiter
$or backtick characters in messages accidentally triggering command substitution
These errors are often silent: the command appears to succeed, but the result is wrong (e.g., a commit message of "fix" instead of "fix: resolve null pointer in auth handler"). LLMs typically cannot observe the actual error because the shell may not surface it clearly in the tool output.
The git-mcp server solves all of this. Every tool parameter is a typed, validated field that the server passes directly to simple-git as an argument array, bypassing shell interpretation entirely. There are no quoting problems, no interpolation hazards, and no silent truncation.
Rule: Always use git-mcp MCP tools for git operations. Never invoke git via shell unless the specific operation is genuinely not available through any git-mcp tool.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 2mo ago First seen · 219 lines · 160 tokens per session scan A b46af2d2ba0f
git-mcp-workflow is a skill published in the GitHub repository selfagency/beans-mcp (0 stars, last pushed 6mo ago), licensed MIT. It adds 160 tokens to every session and 2,683 once invoked, about $0.0006 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 2026-08-31.
git-mcp-workflow compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| git-mcp-workflowselfagency/beans-mcp | Skill | 0 | 6mo ago | MIT | A | 160 | Claude Code, Codex |
| githubinitializ/forge | Skill | 235 | yesterday | Apache-2.0 | A | 15 | Claude Code, Codex |
| issue-triagenearform/lastlight | Skill | 44 | 2d ago | MIT | A | 46 | Claude Code, Codex |
| document-releaseDollarDill/beads-superpowers | Skill | 32 | 7d ago | MIT | A | 103 | Claude Code |
| finishing-a-development-branchDollarDill/beads-superpowers | Skill | 32 | 7d ago | MIT | A | 44 | Claude Code |
| issue-commentnearform/lastlight | Skill | 44 | 2d ago | MIT | A | 47 | Claude Code, Codex |
| surface-after-careellmos-ai/skills | Skill | 9 | 4d ago | MIT | A | 283 | Claude Code, Codex |