Skip to content
Back to search

mcp-snitch

MCP Snitch is a macOS application that intercepts and monitors MCP server communications, providing security analysis, access control, and audit logging for AI tool usage.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/adversis--mcp-snitch/badge.svg)](/listings/adversis--mcp-snitch)

HTML

<a href="/listings/adversis--mcp-snitch"><img src="/api/listings/adversis--mcp-snitch/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

For Users

1. **Download** the latest DMG from [Releases](../../releases) 2. **Install** MCP Snitch by dragging to Applications 3. **Launch** and grant necessary permissions 4. **Protect** your MCP servers from the main interface

For Developers

# Clone the repository
git clone https://github.com/Adversis/mcp-snitch.git
cd mcp-snitch

# Build development version
./build.sh

# Run the app
./run.sh

See [CONTRIBUTING.md](CONTRIBUTING.md) for detailed setup instructions.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 205 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # MCP Snitch
  2. **Security monitoring and access control for Model Context Protocol (MCP) servers**
  3. MCP Snitch is a macOS application that intercepts and monitors MCP server communications, providing security analysis, access control, and audit logging for AI tool usage.
  4. ## Screenshots
  5. ### Server Management
  6. ![Trusted MCP Servers](images/SCR-20251013-hqsz.png)
  7. *View and manage trusted MCP servers with automatic discovery of Claude Desktop and Cursor configurations*
  8. ### Automatic Server Discovery
  9. ![Scan Complete](images/SCR-20251013-hqvr.png)
  10. *Automatically scan and discover MCP servers from your AI tool configurations*
  11. ### Tool Approval System
  12. ![Tool Approval Required](images/SCR-20251013-hwxb.png)
  13. *Approve or block new tool calls with granular control - once, always, or never*
  14. ### Security Blocked Request
  15. ![Request Blocked](images/SCR-20251013-hxcz.png)
  16. *MCP Snitch blocks suspicious requests and provides clear feedback*
  17. ## Features
  18. ### 🔒 Security Monitoring
  19. - **Real-time interception** of stdio and HTTP MCP transports
  20. - **AI-powered security analysis** using GPT-3.5 for threat detection
  21. - **Pattern-based detection** for sensitive data (SSH keys, credentials, system files)
  22. - **Audit logging** with full request/response history
  23. ![Audit Log](images/SCR-20251013-hrja.png)
  24. *Comprehensive audit log showing all MCP tool calls with risk assessment and filtering*
  25. ### 🛡️ GuardRails
  26. - **Approve/Block mode** - Manual approval for tool calls
  27. - **Auto-approve trusted tools** - Whitelist safe operations
  28. - **Trust management** - Per-server trust settings with database storage
  29. - **Risk assessment** - AI analysis of tool call intentions
  30. ![GuardRails Configuration](images/SCR-20251013-hqyx.png)
  31. *Configure security policies including tool control modes and threat detection*
  32. ![AI Analysis Settings](images/SCR-20251013-hrbn.png)
  33. *Enable AI-powered security analysis with Claude or OpenAI*
  34. ![Custom Security Prompts](images/SCR-20251013-hrds.png)
  35. *Customize the AI security analysis prompt for your specific needs*
  36. ![Whitelisted Tools](images/SCR-20251013-hrlj.png)
  37. *Manage whitelisted and blacklisted tools for automated control*
  38. ### 🌐 Universal MCP Support
  39. - **Stdio servers** - npx, Docker, local binaries
  40. - **HTTP servers** - Remote MCP endpoints (GitHub MCP, etc.)
  41. - **Session management** - Automatic session ID handling
  42. - **Keychain integration** - Secure API key storage
  43. ![API Keys Management](images/SCR-20251013-hrsg.png)
  44. *Protect API keys by moving them to macOS Keychain for secure storage*
  45. ### 📊 Monitoring
  46. - **Live message viewer** - See all MCP communications in real-time
  47. - **Server discovery** - Automatic detection of Claude Desktop & Cursor configs
  48. - **Protection status** - Visual indicators for monitored servers
  49. - **Export logs** - Full audit trail for compliance
  50. ## Quick Start
  51. ### For Users
  52. 1. **Download** the latest DMG from [Releases](../../releases)
  53. 2. **Install** MCP Snitch by dragging to Applications
  54. 3. **Launch** and grant necessary permissions
  55. 4. **Protect** your MCP servers from the main interface
  56. ### For Developers
  57. ```bash
  58. # Clone the repository
  59. git clone https://github.com/Adversis/mcp-snitch.git
  60. cd mcp-snitch
  61. # Build development version
  62. ./build.sh
  63. # Run the app
  64. ./run.sh
  65. ```
  66. See [CONTRIBUTING.md](CONTRIBUTING.md) for detailed setup instructions.
  67. ## How It Works
  68. MCP Snitch acts as a transparent proxy between AI applications (Claude Desktop, Cursor) and MCP servers:
  69. ```
  70. Cursor → MCP Snitch Proxy → MCP Server
  71. ↓
  72. Security Analysis
  73. Audit Logging
  74. GuardRails Enforcement
  75. ```
  76. ### Stdio Proxy
  77. - Wraps command-based servers (npx, docker, local binaries)
  78. - Intercepts stdin/stdout for JSON-RPC analysis
  79. - Transparent to both client and server
  80. ### HTTP Proxy
  81. - Local HTTP proxy for remote MCP servers
  82. - Manages session IDs automatically
  83. - Injects authentication headers from keychain
  84. ## Security Analysis
  85. MCP Snitch uses multiple layers of security analysis:
  86. 1. **Pattern Detection** - Fast regex-based checks for known threats
  87. 2. **Trust Verification** - Database lookup for trusted servers
  88. 3. **AI Analysis** - GPT-3.5 analysis for complex tool calls
  89. 4. **Response Monitoring** - Scans outputs for sensitive data leaks
  90. ### Example Threats Detected
  91. - File system access to sensitive directories (`/etc`, `/System`)
  92. - SSH key or credential exposure
  93. - Unusual API calls or data exfiltration patterns
  94. - System command execution
  95. ## Configuration
  96. ### Trusted Servers
  97. Trust servers to skip approval prompts:
  98. - Settings → Servers → Trust/Untrust
  99. ### GuardRails Modes
  100. - **Off** - No interception
  101. - **Approve** - Manual approval for all tool calls
  102. - **Block** - Automatic blocking with AI analysis
  103. ### API Keys
  104. Store OpenAI API key securely in macOS Keychain:
  105. - Settings → API Keys → Add Key
  106. ## Architecture
  107. ```
  108. MCPSnitch/ # Main macOS app (SwiftUI)
  109. ├── Views/ # UI components
  110. ├── Models/ # Data models & database
  111. └── Managers/ # Business logic
  112. MCPProxy/ # Proxy implementations
  113. ├── mcp_proxy.swift # Stdio proxy
  114. ├── mcp_http_proxy.swift # HTTP proxy
  115. └── mcp_security_common.swift # Shared security logic
  116. ```
  117. ## Requirements
  118. - macOS 13.0 or later
  119. - Xcode 15.0+ (for building from source)
  120. - OpenAI API key (for AI security analysis)
  121. ## Licensing
  122. This project is available under a dual-license model:
  123. - **Open Source License**: GNU General Public License v3.0 (GPL-3.0) - See [LICENSE](LICENSE) file
  124. - **Commercial License**: For proprietary use without GPL obligations - Contact licensing@adversis.io
  125. ### Why Dual Licensing?
  126. - **Open Source**: Free use under GPL-3.0 for open source projects and personal use
  127. - **Commercial**: Avoid GPL obligations for proprietary/commercial software
  128. For commercial licensing inquiries, please contact licensing@adversis.io
  129. ## ⚠️ DISCLAIMER
  130. THIS SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  131. **SECURITY NOTICE**: This tool is designed for authorized security monitoring only. Users are responsible for complying with all applicable laws and regulations. Unauthorized use of this tool may result in criminal and/or civil penalties. Always ensure you have proper authorization before monitoring any systems or applications.
  132. ## Contributing
  133. We welcome contributions! See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
  134. ## Support
  135. - **Issues** - Report bugs or request features via [GitHub Issues](../../issues)
  136. - **Discussions** - Ask questions in [GitHub Discussions](../../discussions)
  137. ## Acknowledgments
  138. Built with:
  139. - Swift & SwiftUI
  140. - SQLite for trust database
  141. - OpenAI GPT-3.5 for security analysis
  142. - Model Context Protocol specification
  143. ---
  144. Copyright © 2025 Adversis, LLC. Licensed under GPL-3.0 or Commercial License.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.