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/layered-ai-public/claude-review-actionWrote 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/commands/layered-ai-public/claude-review-action/review-branch-and-fix)<a href="https://agentmods.dev/commands/layered-ai-public/claude-review-action/review-branch-and-fix"><img src="https://agentmods.dev/badge/commands/layered-ai-public/claude-review-action/review-branch-and-fix/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/commands/layered-ai-public/claude-review-action/review-branch-and-fix"><img src="https://agentmods.dev/badge/commands/layered-ai-public/claude-review-action/review-branch-and-fix.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 | $0.00011 | $0.00657 |
| Opus 5.5 | $0.00004 | $0.00263 |
| Sonnet 5.5 | $0.00002 | $0.00131 |
| Haiku 4.5 | $0.00001 | $0.00066 |
Grade A, and why
review-branch-and-fix 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 12d 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 — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review and fix the current branch against the base branch.
The review itself is defined by the /review-branch command - this command does not restate it. Run that review, then fix what it finds.
Base branch
Determine the base branch by checking which of main or master exists as a remote tracking branch. If the argument $ARGUMENTS is provided, use that as the base instead.
Run git diff <base>...HEAD to obtain the diff. If there are no changes, say so and stop.
Running a review pass
Each review pass is delegated to a subagent running Sonnet, so the review is done by a different model from the one applying the fixes.
Use the Agent tool with subagent_type: "general-purpose" and model: "sonnet", and this prompt:
Invoke the
review-branchcommand with the Skill tool, passing<base>as its argument, and follow its instructions exactly. Do not modify any files. Return the complete review as your final message, in the output format that command specifies, and nothing else. If noreview-branchcommand is available to you, say so instead of reviewing from your own judgement.
Substitute the base branch you determined above. Print the review the subagent returns verbatim, labelled with the pass number (e.g. "Review pass 1").
The review-branch command is installed alongside this one, so refer to it by name and let Claude Code resolve it - do not hardcode a path to the file. If the subagent reports it cannot find the command, stop and say that review-branch needs to be installed; do not fall back to reviewing from memory or from these instructions, which contain no review rules.
If the Agent tool is unavailable, invoke review-branch yourself in this session for each pass and note that it ran on this session's model rather than Sonnet.
Fix cycle
Run review pass 1. Then, if there are any MEDIUM or above issues:
- Fix each issue in the code. Keep changes minimal and focused - only fix what the review identified. Do not refactor, reorganise, or "improve" surrounding code.
- Never apply a fix based on an assumption the review could not verify. If a finding rests on the review's "Assumptions" line, verify it yourself in the codebase first; if you cannot, leave it unfixed and say why.
- After applying all fixes, run the next review pass the same way. Report only new issues introduced by your fixes - do not re-report issues that have already been resolved.
- If new issues are found, fix them and review again.
- Repeat until a clean review pass with no MEDIUM or above issues, or you have completed 3 fix cycles (whichever comes first).
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.
- 12d ago First seen · 45 lines · 11 tokens per session scan A 65a178ac7fa1
review-branch-and-fix is a command published in the GitHub repository layered-ai-public/claude-review-action (11 stars, last pushed 7d ago), licensed Apache-2.0. It adds 11 tokens to every session and 657 once invoked, about $0.0000 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-19.
Other commands, from other repositories
speckit.spex.submit
Push and create PR for team review, with optional watch mode for CI monitoring.
git
The pre-finish status: branch, hygiene findings, message checks, workflow lint, template state.
merge-conflict-analysis
You are analyzing merge conflicts for PR #${{ pr-number }}.
factory
Linha de montagem multi-agente (Software Factory) — spec-writer → builder em worktree isolado → 4 revisores paralelos (Security, Code, UI, UX) → approver com gate de risco e dashboard visual.
repo-audit
Audit a codebase (local or remote GitHub/GitLab) against architecture principles and requirements, surfacing drift, risk, and missing decisions.
release-tag
Create and push a version tag after a release PR has been merged.