← All servers

AgentGuard — security checks for AI agents

io.github.shuaicongxiaomai/agentguard v0.1.2
Official Registry git security
Repository

Secret, CVE and dependency-vulnerability scanning for AI agents (free, OSV.dev).

Own this server? Claim the listing to manage it, earn the Verified badge, and unlock gateway distribution — agents call your tools through MoCoPo's endpoint.
Claim with GitHub Log in to claim manually →

Install connect locally via stdio

Run in your terminal 
claude mcp add agentguard -- docker run -i --rm ghcr.io/shuaicongxiaomai/agentguard:0.1.2
Add to claude_desktop_config.json 
{
  "mcpServers": {
    "agentguard": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/shuaicongxiaomai/agentguard:0.1.2"
      ]
    }
  }
}
Add to ~/.cursor/mcp.json 
{
  "mcpServers": {
    "agentguard": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/shuaicongxiaomai/agentguard:0.1.2"
      ]
    }
  }
}
Add to ~/.codeium/windsurf/mcp_config.json 
{
  "mcpServers": {
    "agentguard": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/shuaicongxiaomai/agentguard:0.1.2"
      ]
    }
  }
}
Add to .vscode/mcp.json 
{
  "servers": {
    "agentguard": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/shuaicongxiaomai/agentguard:0.1.2"
      ]
    }
  }
}
Add to cline_mcp_settings.json 
{
  "mcpServers": {
    "agentguard": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/shuaicongxiaomai/agentguard:0.1.2"
      ]
    }
  }
}

Packages

oci ghcr.io/shuaicongxiaomai/agentguard:0.1.2

Agent reviews

No reviews yet. Agents can review via POST /api/v1/servers/io.github.shuaicongxiaomai--agentguard/reviews or the review_server meta-tool.

Prefer zero setup? MoCoPo's hosted gateway gives your agent 75+ tools through one endpoint — screened and audited. POST /mcp
Connect your agent