ReviewEngine CLI
CLI with a virtual board of AI experts that reviews a PR or local diff and returns a score, severity level and evidence.
- 104 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Code changes are sent to an external AI provider.
- With a token and `--publish` the tool posts to PR and MR discussions as the token owner.
Install
- Works in
Manual install
curl -fsSL https://raw.githubusercontent.com/Liewzheng/ReviewEngine/master/install.sh | bashThe first README command installs the binary and the `reng` alias. You can download and inspect the script first.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Code changes are sent to an external AI provider.
- With a token and `--publish` the tool posts to PR and MR discussions as the token owner.
This is third-party code. Review the repository files before installing.
What it does
ReviewEngine is a static Rust binary that runs a board of AI experts: security, performance, quality, code reuse and documentation. The experts look at the same diff in parallel, each through its own lens. The output is a Markdown or JSON report. Every finding carries severity, confidence, evidence, impact, a recommendation and effort, and the report adds per-expert scores, a weighted overall score and a risk level from low to critical. It can review a local repository (`--base`, `--staged`, `--since`), a GitHub PR or a GitLab MR, and the `--publish` flag posts findings to the discussion. The `review-engine` skill launches the CLI from Claude Code on phrases like "review a PR".
Who it is for. Teams that want one scored review format, and developers who check a branch before opening a PR.
Good fit when
- Before creating a PR the branch must be compared with `main` and high-severity findings fixed.
- A huge diff needs a map of security and performance concerns for the reviewer.
- A GitLab or GitHub CI job needs a step that writes a structured report to the MR discussion.
- A Claude Code agent should launch the expert board on the phrase "review this repo".
Not a fit when
- Code fragments must not be sent to an AI provider.
- A plain review request without the tool name is needed; the skill intentionally ignores it.
- SSO, audit logs or custom expert templates are needed without a commercial license.
Example request
Run review-engine on the current branch against `main` and show the high-severity findings.Limitations
An AI provider key is required (OpenAI, Anthropic, DeepSeek or a compatible one), set in `.code-audit-config.toml` or the `LLM_CONFIG` variable. Code changes are sent to the provider. PRs and MRs need `GITHUB_TOKEN` or `GITLAB_TOKEN`, and with `--publish` the report lands in the discussion. The installer needs `curl`, `jq` and `sha256sum` (`shasum` on macOS). SSO, audit logs and custom expert templates belong to a separate commercial edition. A model assigns the scores, so verify findings by hand.
How to disable. Delete the `review-engine` directory from `~/.claude/skills/` and the `review-engine` and `reng` binaries. If you ran `reng serve` or webhooks, stop the server and revoke the GitHub and GitLab tokens.
FAQ
How do I review a branch before a PR?
Run `reng review` with `--local-path` and `--base main`. Use `--staged` for the index and `--since` with `--until` for a commit range.
How do I post the report to an MR?
Add `--publish` and set `GITLAB_TOKEN` or `GITHUB_TOKEN` in the environment. Avoid passing tokens as flags because they end up in shell history.
Which models work?
OpenAI, Anthropic, DeepSeek and any OpenAI-compatible server. Set the provider in `.code-audit-config.toml` or through `LLM_CONFIG`.
How do I cut noise from findings?
Raise `min_confidence` and enable `drop_low_confidence` in the `[report]` section. Findings outside the changed lines are dropped automatically.
Is the tool free?
The core, local review, integrations and REST API are free under Apache 2.0. SSO and audit logs are sold separately.
Discussion
No questions or reviews yet.
Related
CLI toolsreviewd reviews GitHub, GitLab and BitBucket pull requests through Claude Code, Gemini or Codex CLI and comments on them
Jev-Code-Reviewer ranks changes in an agent PR as P0, P1 or P2 and retells them in plain language on the GitHub page.
The CLI matches linter diagnostics to code changes and reports findings locally or through an approved review channel.
A code-review CLI and plugin for agents: reads a git diff and writes line-level comments with context
A CLI and three skills that work through review-bot comments on a PR: assess findings, fix code, reply and wait for checks
Two skills: a pull request review where two models argue, and a helper that works through review comments round by round
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| ReviewEngine (this card) | no | 104 | |
| code-review-skill | no | 2,256 | single-agent code review; ReviewEngine runs a weighted expert board with an overall score |
| open-code-review | no | 50,109 | rule-based code checking; here an expert board scores the diff and publishes the report to a PR or MR |
README in short
A single binary with the `reng` alias installs by script or through Homebrew and runs on a laptop, a server or CI. A REST API and GitLab and GitHub webhooks are also available.