🧭 Vibe Check MCP

What is 🧭 Vibe Check MCP?

🧭 Vibe Check MCP is an AI coding safety net that detects engineering anti‑patterns in AI‑generated code before they become expensive mistakes. It integrates with Claude Code to provide real‑time engineering coaching, specifically designed for “vibe coders” who want to avoid overengineering traps, security blindspots, and doom loops.

How to use 🧭 Vibe Check MCP?

Install via one‑command curl script or manual git clone and pip install. Add the server to Claude Code with claude mcp add-json (never use -s user flag). Use natural‑language commands like “quick vibe check issue 42” or “deep vibe check issue 42 with full Claude analysis” to call tools such as analyze_github_issue, analyze_pull_request, analyze_text, analyze_code, and validate_integration. Optionally set a GITHUB_TOKEN for private repository access.

Key features of 🧭 Vibe Check MCP

• Two modes: fast pattern detection and deep Claude‑powered analysis
• Detects 4 core anti‑patterns: infrastructure‑without‑implementation, symptom‑driven development, complexity escalation, documentation neglect
• Validated 87.5% detection accuracy with 0% false positives
• Real‑time educational coaching with case studies
• Seamless integration with Claude Code MCP workflow
• Docker and bridge deployment options available

Use cases of 🧭 Vibe Check MCP

• Catch costly overengineering traps during the planning phase of an AI‑coded project
• Analyze GitHub issues for anti‑patterns before implementation begins
• Review pull requests for unnecessary complexity and architectural problems
• Validate integration approaches (e.g., use official SDK instead of custom HTTP client)
• Get educational feedback on why a pattern is problematic and how to fix it

FAQ from 🧭 Vibe Check MCP

What prerequisites are needed?

Claude Code installed and configured, Python 3.8+ with pip, and optionally a GitHub token with repo and read:org scopes for private repository access.

How does fast analysis differ from deep analysis?

Fast analysis uses local pattern detection without external API calls and responds in under 30 seconds. Deep analysis invokes Claude for comprehensive anti‑pattern detection, educational explanations, and real‑world case studies, taking up to 90 seconds.

Why should I never use the -s user flag when adding the server?

Using -s user causes infinite recursion and Claude Code timeouts. This server only works in Claude Code SDK (non‑interactive) mode.

Can I use 🧭 Vibe Check MCP with Docker?

Yes, Docker deployment is supported. It containers all dependencies and provides isolation; see the MCP Deployment Guide for setup.

What GitHub integration is supported?

The server can analyze GitHub issues and pull requests. Public repositories work automatically; private repos require a valid GITHUB_TOKEN environment variable or GitHub CLI authentication with appropriate permissions.