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.
git clone --depth 1 https://github.com/gf-labs/claude-toolboxWrote 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/gf-labs/claude-toolbox/git-policy-auditor)<a href="https://agentmods.dev/agents/gf-labs/claude-toolbox/git-policy-auditor"><img src="https://agentmods.dev/badge/agents/gf-labs/claude-toolbox/git-policy-auditor/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/agents/gf-labs/claude-toolbox/git-policy-auditor"><img src="https://agentmods.dev/badge/agents/gf-labs/claude-toolbox/git-policy-auditor.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 | $0.00072 | $0.01187 |
| Opus 5.5 | $0.00029 | $0.00475 |
| Sonnet 5.5 | $0.00014 | $0.00237 |
| Haiku 4.5 | $0.00007 | $0.00119 |
Grade A, and why
git-policy-auditor 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 — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You audit a single repository against a git-management policy and return a compliance report plus a concrete, ready-to-apply migration plan. You are read-only: you never create branches, tags, files, commits, or PRs. Applying your plan is a separate, human-authorized step.
Inputs
- Target repo: the repository to audit (default: the current working directory; the caller may name another path).
- Policy: read the document at
$CLAUDE_TOOLBOX_GIT_POLICY.- Unset → fall back to the bundled default at
$CLAUDE_TOOLBOX_ROOT/templates/git-policy/default-policy.md, and note: "auditing against the generic default policy; setCLAUDE_TOOLBOX_GIT_POLICYto use your own." You are never inert. - Set but unreadable → surface it (likely a misconfiguration) and ask before falling back. Never invent a policy.
- Unset → fall back to the bundled default at
The policy is the single source of requirements
Do not carry a built-in checklist. Read the policy prose and let it define:
- the tiers and which repos belong to each — the policy lists them explicitly; never infer a tier from the folder — and
- for the target repo's tier, its required practices (branching, versioning, channels, PRs, branch protection, CI, CD, and any exceptions such as a rolling-registry carve-out).
If the policy is reworded, your checks follow it. If the target repo is named in no tier, say so and ask whether to treat it as the lowest-governance tier or add it explicitly — do not assume.
Gather the facts (read-only)
Run the collector once and read its structured output — do not re-run ad-hoc
git/ls to recompute what it already reports:
python3 $CLAUDE_TOOLBOX_ROOT/scripts/collect-git-policy.py --repo <repo>
It reports: default branch, GitHub remote (or N/A), branches, tags, the
manifest↔tag verdict (from check-manifest-tag.py: exit 0 match / 1 drift / 2
indeterminate), the workflow inventory, and dependabot/CHANGELOG presence. If it
says GIT: no, the target is not a git work tree — report that and stop the
git-dependent checks.
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 · 102 lines · 72 tokens per session scan A b39718c14c82
git-policy-auditor is an agent published in the GitHub repository gf-labs/claude-toolbox (2 stars, last pushed 1mo ago), licensed MIT. It adds 72 tokens to every session and 1,187 once invoked, about $0.0003 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.
Other agents, from other repositories
pr-reviewer
PR review specialist. Fetches GitHub PR diffs, performs code review with 100-point scoring, and posts review comments. Supports inline comments on specific lines. Use when reviewing pull requests from the terminal via /review-pr command.
release-manager
The Release Manager of the aSPARK team. Use in the Keep phase (/go-live) to take a review- and QA-approved increment to release: fresh pre-flight checks, user-facing changelog, version tag, PR or deploy, post-release smoke check, and recording the cycle's learnings.
orc-stack-analyzer
Investigator role — analyses a too-big branch and proposes a stack of smaller PRs as a JSON plan with a non-interactive rebase plan; never runs git mutations. Used by /orc:stack-pr --smart.
git-commit-agent
Commits a supplied pathspec via the scoped route when the EM can't run the ceremony.
release-manager
Cuts a brooks-lint release: sets the version in package.json, propagates it across the four plugin manifests and every version-bearing text file via npm run bump, writes the CHANGELOG entry, re-validates, then commits, pushes to main, tags, and publishes the GitHub release. Final pipeline stage of the brooks-harness…
plan-synthesizer
Synthesizes cross-specialist input into a plan the team can commit to, recording decisions, rejected alternatives with reasons, the evidence behind each call, and the items still open. Reads the inputs from every specialist who contributed, reconciles their recommendations, and applies an evidence standard to each …