Skip to content
MCP server

Security Engineer

Turns GuardBee security findings into reviewable fixes your coding agent applies

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
CLI, API
Runs
Locally
Version
0.2.0
Listing updated
today

Similar MCP servers

About Security Engineer

Security Engineer is an open-source MCP server for security. It runs locally on your machine and is started by your AI client.

GuardBee's own assessment finds and explains a problem; this server hands the fix to the coding agent already working in your repository. The README is clear about the split of duties: GuardBee proposes a fix for your platform, the agent makes the edit, you approve it, and GuardBee never edits your code or servers.

A typical run starts by listing recent assessments, pulling the highest-priority findings and detecting the stack from repository files such as vercel.json, netlify.toml, public/_headers and Dockerfiles that mention nginx. For Nginx or Apache you can pass your config and receive an exact diff. After the change is committed and deployed, the agent records the commit SHA or PR URL and asks GuardBee to re-run the original check against the live site.

The server reads only files inside the project root and only the config file you name, and every action is logged in GuardBee's activity log as coming from an MCP client. Telemetry is off unless you set GUARDBEE_TELEMETRY=1.

Install

Run in a terminal

claude mcp add security-engineer --env GUARDBEE_API_KEY='<GUARDBEE_API_KEY>' -- npx -y @guardbee/mcp-security-engineer

Add to claude_desktop_config.json

{
  "mcpServers": {
    "security-engineer": {
      "command": "npx",
      "args": [
        "-y",
        "@guardbee/mcp-security-engineer"
      ],
      "env": {
        "GUARDBEE_API_KEY": "<GUARDBEE_API_KEY>"
      }
    }
  }
}

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "security-engineer": {
      "command": "npx",
      "args": [
        "-y",
        "@guardbee/mcp-security-engineer"
      ],
      "env": {
        "GUARDBEE_API_KEY": "<GUARDBEE_API_KEY>"
      }
    }
  }
}

Add to .vscode/mcp.json

{
  "servers": {
    "security-engineer": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@guardbee/mcp-security-engineer"
      ],
      "env": {
        "GUARDBEE_API_KEY": "<GUARDBEE_API_KEY>"
      }
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "security-engineer": {
      "command": "npx",
      "args": [
        "-y",
        "@guardbee/mcp-security-engineer"
      ],
      "env": {
        "GUARDBEE_API_KEY": "<GUARDBEE_API_KEY>"
      }
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "security-engineer": {
      "command": "npx",
      "args": [
        "-y",
        "@guardbee/mcp-security-engineer"
      ],
      "env": {
        "GUARDBEE_API_KEY": "<GUARDBEE_API_KEY>"
      }
    }
  }
}

Using several servers? Build one config file with the MCP config generator.

Configuration

  • GUARDBEE_API_KEYGuardBee API key with remediations.read and remediations.write scopesRequired

Tools this server provides (6)

Taken from the project documentation; the running server may differ.

list_assessments
Recent AI Security Engineer assessments of your workspace.
get_assessment
Actionable findings, highest priority first.
detect_stack
Detects the stack from repository files.
get_fix
A reviewable fix for a finding, with a diff when a config is supplied.
report_applied
Records that the fix was applied.
verify_fix
Re-runs the check against the live site.

What it does

Prioritised findings
Lists actionable findings, highest priority first, with the platforms each can be fixed on.
Stack detection
Detects Next.js, Express, Nginx, Apache or Cloudflare from repository files.
Reviewable fixes
Returns a fix per finding, with an exact diff when you supply an Nginx or Apache config.
Applied-fix record
Stores the commit SHA or PR URL against the finding.
Live verification
Re-runs the original check on the deployed site.
Project-root sandbox
Rejects paths outside the project root and reads only the named config file.

Good for

  • Work through the top findings of the latest GuardBee assessment
  • Patch security headers in a Next.js or Nginx setup
  • Confirm on the live site that a deployed fix worked
  • Keep an audit trail of agent-made security changes

Pricing

Open source

MIT licensed; needs a GuardBee API key with remediations scopes

Security Engineer FAQ

Does GuardBee change my code?

No. Your coding agent makes the edits and you approve them.

What credentials are needed?

A GuardBee API key with the remediations.read and remediations.write scopes, passed as GUARDBEE_API_KEY.

Which stacks can it detect?

Next.js, Express, Nginx, Apache and Cloudflare, plus a note when Fly or Firebase hosting is present.

Is telemetry collected?

Not by default; it is switched on only with GUARDBEE_TELEMETRY=1.

Is Security Engineer free?

Yes. Security Engineer is open source, so you can use and self-host it for free.

How do I add Security Engineer to Claude Code?

Run this command in your terminal: claude mcp add security-engineer --env GUARDBEE_API_KEY='<GUARDBEE_API_KEY>' -- npx -y @guardbee/mcp-security-engineer

What are the best alternatives to Security Engineer?

The closest alternatives to Security Engineer are REA, Context7, Chrome DevTools MCP, Codebase Memory and GitHub.

Where this comes from

  • Description: written from the project's README.
  • Install steps: from the official MCP Registry entry.

How AIPoweredBot checks servers