MCP Defender
App : A product you use directly
Desktop app that intercepts and filters MCP traffic from installed AI clients.
Interesting idea for a workstation, still a young product.
Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.
What it does, concretely
AI applications like Cursor, Claude or Visual Studio Code communicate with MCP (Model Context Protocol) servers to run tools. These exchanges can be hijacked through attacks such as prompt injection, SSH key theft or malicious code execution without the user's knowledge.
MCP Defender acts as a secure proxy between your AI applications and MCP servers. It analyses every tool call in real time using a combination of deterministic signatures and large language model analysis, then alerts or blocks any suspicious activity before it can cause damage.
The tool is aimed at developers who use AI clients on their workstation and want to keep control over what those applications actually do in the background. It supports Cursor, Claude, Visual Studio Code and Windsurf, on macOS and Windows.
It is less suited to users who do not run desktop AI clients with MCP, or to those looking for a broader network security solution: MCP Defender focuses exclusively on local MCP traffic. Note that the project has been acquired by Docker Inc., which may influence its future direction.
What it offers
- ·Automatic proxy that intercepts all MCP traffic between your AI applications and connected servers.
- ·Threat detection combining LLM analysis and deterministic signatures to identify malicious activity.
- ·Protection against prompt injection, credential theft, arbitrary code execution and remote command injection.
- ·Real-time alerts as soon as suspicious activity is detected, without slowing down monitored applications.
- ·Signature management, customisable to adapt scanning rules to your needs.
- ·Choice of LLM provider and use of your own API keys for threat analysis.
- ·Open source code under the AGPL-3.0 licence, available on GitHub.
- ·Compatible with Cursor, Claude, Visual Studio Code and Windsurf on macOS and Windows.
In practice
- Price
- Freemium
- Commercial use
- Commercial use: yes
- Available on
- macos, windows
Security and data
- Rating
- Security: one caveat, see the card
- Your data
- Data stays local
- Maintenance
- Active
Links and repository checked on 25 October 2026 · added on 1 June 2025