MCP Code Review

MCP Code Review Server is an open-source MIT-licensed code review tool delivered as a Python package (`aicraft-code-review`). It runs locally and connects to MCP-compatible AI editors such as Claude Code, Cursor, and Windsurf. The server performs static analysis and AI-assisted review, identifying security issues (OWASP vulnerabilities, SQL injection, command injection), performance problems (N+1 queries), race conditions, and other defects, with severity ratings and suggested fixes. The tool outputs a structured review report containing high, medium, and info severity findings, including line numbers, categories, and specific fixes. A local configuration file (`.mcp-code-review.yaml`) supports custom rules and team profiles, enabling consistent review behavior across team members and machines. The free version is available via `pip install aicraft-code-review` and has been listed on Smithery, mcpservers.org, and cursor.directory. An optional $49 Team Rules Pack adds 63 validated rules for Python, JavaScript/TypeScript, Go, and Java, CI workflows that block merges on critical findings, and 20 LLM review prompts. The tool does not upload code to external services, addressing privacy and security concerns for teams that review proprietary source code.

Category: Apis
Pricing: Paid

Problem solved

Automates local-first AI code review inside MCP-compatible editors, catching security, performance, and correctness issues before code is merged, while keeping all code on the user's machine and allowing team-shared rule configurations.

Use cases

Scan a Python file before commit for security and performance issues
Set up team-shared code review rules across Claude Code, Cursor, and Windsurf
Block merges in CI when critical findings are detected
Detect OWASP vulnerabilities, N+1 queries, SQL injection, and race conditions
Integrate with any MCP-compatible AI editor for inline review feedback
Use prebuilt rule packs for Python, JavaScript/TypeScript, Go, and Java
Run local code reviews without uploading source code to external services

Audience

Developers
ML Engineers
Data Scientists
Startups
Enterprise

Pros

Runs locally and never uploads code
Open-source MIT and free via pip
Detects OWASP security, N+1, SQL injection, and race conditions
Supports custom rules and team profiles via YAML
Works with Claude Code, Cursor, Windsurf, and other MCP editors

Cons

Requires a Python environment and pip installation
Advanced team rules pack costs $49
No built-in web dashboard or GUI
Free version may lack prebuilt rules for some languages
Requires an MCP-compatible editor for interactive use

Official website: https://aicraft.vip/

Canonical LambdaBase page: https://www.lambdabase.com/tools/mcp-code-review

← Directory

MCP Code Review

Paid

★ 1 • Updated Aug 16, 2026, 09:17 AM

MCP Code Review Server is an open-source MIT-licensed code review tool delivered as a Python package (`aicraft-code-review`).

Use Cases ▼

Scan a Python file before commit for security and performance issues

Set up team-shared code review rules across Claude Code, Cursor, and Windsurf

Block merges in CI when critical findings are detected

Detect OWASP vulnerabilities, N+1 queries, SQL injection, and race conditions

Integrate with any MCP-compatible AI editor for inline review feedback

Use prebuilt rule packs for Python, JavaScript/TypeScript, Go, and Java

Run local code reviews without uploading source code to external services

What Problem It Solves ▼

Key Features ▼

✓

Runs locally and never uploads code

✓

Open-source MIT and free via pip

✓

Detects OWASP security, N+1, SQL injection, and race conditions

✓

Supports custom rules and team profiles via YAML

✓

Works with Claude Code, Cursor, Windsurf, and other MCP editors

Overview

Full Description ▼

The tool outputs a structured review report containing high, medium, and info severity findings, including line numbers, categories, and specific fixes. A local configuration file (`.mcp-code-review.yaml`) supports custom rules and team profiles, enabling consistent review behavior across team members and machines. The free version is available via `pip install aicraft-code-review` and has been listed on Smithery, mcpservers.org, and cursor.directory.

An optional $49 Team Rules Pack adds 63 validated rules for Python, JavaScript/TypeScript, Go, and Java, CI workflows that block merges on critical findings, and 20 LLM review prompts. The tool does not upload code to external services, addressing privacy and security concerns for teams that review proprietary source code.