Governance proxy that filters MCP tool results for prompt injection and capability abuse using configurable policy modes.
Copy the AI prompt to install this server into Claude Code, Cursor, or another agent โ or use 1-click editor setup below.
We haven't yet run this listing's install command through our automated sandbox check. This isn't a red flag โ we're steadily working through the catalog.
๐ก Paste the JSON block into your client's configuration file under mcpServers, then restart the application.
Inspect callable tools, capabilities, and parameters exposed to AI agents by Arc Gate MCP.
9hannahnine-jpg/arc-gate-mcp MCP server acts as a governance proxy for MCP tool calls. Instead of connecting an MCP client directly to an upstream server, the client connects through Arc Gate. Results returned by the upstream server pass through Arc Gate governance before they are delivered to the agent.
The stated focus is the MCP trust boundary: the proxy is designed to block prompt injection attacks delivered through tool results and to prevent capability abuse. It is therefore suited to deployments where an agent consumes data from tools that may contain untrusted or adversarial content.
Run Arc Gate with an upstream MCP server command and select a policy. The --upstream value identifies the server that Arc Gate should proxy, while --policy selects the enforcement profile. The README demonstrates an upstream Fetch server launched with uvx mcp-server-fetch.
Available policy modes are:
balanced for general-purpose usebrowser_agent for web-browsing agentsfinance_agent for financial-data agentsrag_assistant for document-retrieval agentsstrict for maximum enforcementThe proxy sits in the request and response path; the documented governance behavior applies before tool results reach the agent. The material does not describe the individual detection rules, policy configuration format, or how blocked results are reported.
Install the package from PyPI with:
A direct launch example is:
For Claude Desktop, the README configures the server through uvx, passing arc-gate-mcp, the upstream command, and the browser_agent policy in the MCP server definition. Replace the example upstream or policy with the values appropriate for the MCP server and agent workload you are governing.
The primary capability is proxy-based runtime governance rather than a separate business-data tool set. Supported behavior documented for this project includes:
The available policy names provide workload-oriented choices for browsing, finance, retrieval, general use, and strict enforcement. No additional MCP tools, APIs, credentials, or environment variables are documented.
9hannahnine-jpg/arc-gate-mcp MCP server requires an upstream MCP server command to proxy; the README does not present it as a standalone replacement for an application server. The documentation does not specify whether policies can be customized, how false positives are handled, or what content is logged. It also does not list compatibility details for MCP clients other than the Claude Desktop configuration example.
The project is licensed under AGPL-3.0. The README links to Bendex Arc Platform and Heym, including a governed web research template, but those references do not document additional local configuration or authentication requirements for this package.
Factual signals from GitHub, npm, and our automated checks โ not a rating.
No reviews yet โ be the first to share how this listing worked for you.
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
[](https://allmcps.com/mcp/9hannahnine-jpg-arc-gate-mcp)<a href="https://allmcps.com/mcp/9hannahnine-jpg-arc-gate-mcp"><img src="https://allmcps.com/api/badge/9hannahnine-jpg-arc-gate-mcp?style=directory" alt="Arc Gate MCP on AllMCPs" /></a>