ConnectSecure MCP — Claude Code MCP server: features & setup
Connects the ConnectSecure API to AI agents to query vulnerabilities and CVE exposure across MSP clients.
Open on GitHubWhat ConnectSecure MCP does
This server integrates your AI agent directly with ConnectSecure endpoints. It keeps a local data store, making it easy to correlate security flaws with specific client computers in one step.
The setup script drops two binaries into your PATH: connectsecure-cli and connectsecure-mcp. A crucial security constraint applies here: running with --transport http is unsafe, because this build lacks request size caps and timeouts while holding write permissions.
Stick to the default stdio transport for local agent sessions. Verify the binaries right after installation by executing connectsecure-cli --version in your terminal.
What it helps with
- rank exposure across multiple MSP client environments
- track unpatched CVEs affecting client machines
- identify blind spots across managed endpoints
- trigger ConnectSecure API actions through the AI agent
Example prompts for the agent
- List all critical CVEs affecting client workstations
- Show security exposure rankings across my client base
- Identify endpoints missing recent vulnerability scans
How to install ConnectSecure MCP
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about ConnectSecure MCP
How to install ConnectSecure MCP in Claude Code?
Add the repository using /plugin marketplace add Servosity/msp-skills, then run /plugin install connectsecure@msp-skills. Test the CLI with connectsecure-cli --version.
What does ConnectSecure MCP do?
It gives your assistant access to the ConnectSecure API alongside a local vulnerability cache. You can query threat levels and client assets directly in chat.
Can I run ConnectSecure MCP over HTTP?
Avoid --transport http even on localhost. The current build lacks request timeouts and size limits, meaning stdio is the only safe communication mode.
