Code ReviewSAFE
An MCP server for code reviews using OpenAI and Google models for Claude-code
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Version: 1.0.0
An MCP (Model Context Protocol) server that provides a powerful tool to perform code reviews using various Large Language Models (LLMs). This server is designed to be seamlessly integrated with AI coding assistants like Anthropic's Claude Code, Cursor, Windsurf, or other MCP-compatible clients.
Note: This tool was initially created for Claude Code but has since been expanded to support other AI IDEs and Claude Desktop. See the integration guides for Claude Code, Cursor, and Windsurf.
It analyzes git diff output for staged changes, differences from HEAD, or differences between branches, providing contextualized reviews based on your task description and project details.
Features
- Reviews git diffs (staged changes, current HEAD, branch differences).
- Integrates with Google Gemini, OpenAI, and Anthropic models through the Vercel AI SDK.
- Allows specification of task description, review focus, and overall project context for tailored reviews.
- Outputs reviews in clear, actionable markdown format.
- Designed to be run from the root of any Git repository you wish to analyze.
- Easily installable and runnable via
npxfor immediate use.
Compatibility
- Node.js: Version 18 or higher is required.
- Operating Systems: Works on Windows, macOS, and Linux.
- Git: Version 2.20.0 or higher recommended.
Prerequisites
- Node.js: Version 18 or higher is required.
- Git: Must be installed and accessible in your system's PATH. The server executes
gitcommands. - API Keys for LLMs: You need API keys for the LLM providers you intend to use. These should be set as environment variables:
GOOGLE_API_KEYfor Google models.OPENAI_API_KEYfor OpenAI models.ANTHROPIC_API_KEYfor Anthropic models.
These can be set globally in your environment or, conven
014b452d1e80OBSERVED · 2026-10-08Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add code-review-mcp --env ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} --env GEMINI_API_KEY=${GEMINI_API_KEY} --env GOOGLE_API_KEY=${GOOGLE_API_KEY} --env OPENAI_API_KEY=${OPENAI_API_KEY} -- npx -y @vibesnipe/[email protected]{
"mcpServers": {
"code-review-mcp": {
"command": "npx",
"args": [
"-y",
"@vibesnipe/[email protected]"
],
"env": {
"ANTHROPIC_API_KEY": "${ANTHROPIC_API_KEY}",
"GEMINI_API_KEY": "${GEMINI_API_KEY}",
"GOOGLE_API_KEY": "${GOOGLE_API_KEY}",
"OPENAI_API_KEY": "${OPENAI_API_KEY}"
}
}
}
}Exposed tools (1)
0 read · 1 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
perform_code_review | write | Performs a code review using a specified LLM on git changes. Requires being run from the root of a git repository. |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- declared (3 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (3)
@modelcontextprotocol/sdk, ai, @ai-sdk/openai, @ai-sdk/anthropic, @ai-sdk/google, dotenv, zod, execa
For the LLM integration to work, the `claude-code-review-mcp` server (the process started by `npx` or `claude-code-review-mcp`) needs access to the respective API keys.
The server will automatically load variables from a `.env` file found in the current working directory (i.e., your project's root) or you can configure them directly in the MCP server configuration as
Gates applied: no_behavioural_pass.
014b452d1e80full audit observations/trust-audit/mcp-server/praneybehl__code-review-1.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | 014b452d1e80 | SAFE | B | 89 | first audit |
Questions
What is the Code Review MCP server?
An MCP server for code reviews using OpenAI and Google models for Claude-code
What tools does Code Review expose?
1 in total: 0 read-only, 1 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is Code Review safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does Code Review need?
It reads ANTHROPIC_API_KEY, GEMINI_API_KEY, GOOGLE_API_KEY and OPENAI_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does Code Review run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as @vibesnipe/code-review-mcp at 1.0.0.
How current is this page?
The grade is for one exact copy of the source (014b452d1e80), read on 2026-10-08. The repository is watched and re-audited when it changes.