Robin CLI
Robin reviews every pull request in your repository as a GitHub Action with your own AI key, including OpenRouter's free tier.
- 55 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The diff and files the model reads through tools go to the AI address you configured, and the key is stored in repository secrets.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/antongulin-robinDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The diff and files the model reads through tools go to the AI address you configured, and the key is stored in repository secrets.
This is third-party code. Review the repository files before installing.
What it does
Robin is AI code review for pull requests. A GitHub Actions workflow starts when a request opens or someone comments `/robin`, reads the diff and writes a short summary with inline comments on changed lines and one-click suggested fixes. When nothing is worth flagging it reports "No issues found". There is no separate service with its own servers: the action runs in your repository and calls the address you configure. Installation is one command, `npx robin-review`, from the repository root, or the `install.sh` script without Node.js. The installer creates `.github/workflows/robin.yml` and installs a companion skill for agents so a chat with Claude Code, Cursor or Copilot can add Robin following the rules in `AGENTS.md`. The key and address are set through the `LLM_API_KEY`, `LLM_BASE_URL` and `LLM_MODEL` secrets.
Who it is for. Teams and solo developers on GitHub who want pull request reviews with no service subscription and no per-seat fee.
Good fit when
- Automatic comments on every pull request are wanted with a free OpenRouter key.
- Review must run on a `/robin` comment rather than on each commit.
- An agent must add the reviewer to the repository itself from a ready instruction.
- The model must read callers, types and tests before reporting findings.
Not a fit when
- Code must not go to an external model: the diff and files go to the configured address.
- A weak local model with small context: the tools mode is better switched off in advance.
- A human reviewer is needed: the tool spots individual defects and makes no decisions.
Example request
Add Robin to this repository: create the workflow, tell me which three secrets to create and test it on a sample pull request.Limitations
The repository secrets `LLM_API_KEY`, `LLM_BASE_URL` and `LLM_MODEL` are required; the free option uses an OpenRouter key. Since version 2.13.0 `agent-mode` defaults to `auto`: the model reads callers and tests, which costs more requests, tokens and time, and the files it reads go to your model address. For weak models set `agent-mode: off` in `.github/robin.yml`. Run the installer separately in each repository, not from the home folder. Do not use the `v0` tag or `pull_request_target`.
How to disable. Delete `.github/workflows/robin.yml` and the secrets from repository settings. Remove the Robin skill from the agent skills directory; an archive of older versions sits in `.github/robin-workflow-archive/`.
FAQ
How much does Robin cost?
The tool itself is free under the MIT licence. You pay only the model provider, and OpenRouter has a free tier.
How do I start a review?
Open a pull request or leave a `/robin` comment. The result arrives as a summary and comments on changed lines.
Which mode reads files?
`agent-mode: auto`, on by default. For diff-only review set `agent-mode: off`.
Which secrets are needed?
`LLM_API_KEY`, `LLM_BASE_URL` and `LLM_MODEL`. They go into the GitHub repository secrets.
Can I install without Node.js?
Yes, `curl -fsSL https://robinreview.dev/install.sh | bash` does the same. The script is `scripts/install.sh` and is worth reading before running.
Discussion
No questions or reviews yet.
Related
CLI toolsGitHub Action and CLI: several models review a pull request in parallel and duplicates merge into one report with a cost.
A skill and CLI for code review: a team of reviewer agents inspects changes independently, debates findings and produces one report
A code-review CLI and plugin for agents: reads a git diff and writes line-level comments with context
reviewd reviews GitHub, GitLab and BitBucket pull requests through Claude Code, Gemini or Codex CLI and comments on them
CodeRabbit plugin and skills for Claude Code, Cursor, Codex and 35 more agents: change reviews, fixes from comments, config
A CLI and three skills that work through review-bot comments on a PR: assess findings, fix code, reply and wait for checks
README in short
Describes review features, a prompt for agents, the quick installer, three manual setup steps, agent modes and request settings.