🚀 Start Scanning in 5 Minutes

Secure your AI code with just 3 simple steps

1
Get API Key
2
Add MCP Server
3
Start Scanning
1

Get Your API Key

2 minutes • Authentication required

2

Add MCP Server to Claude Code

2 minutes • Claude Code required

Use Claude Code's built-in MCP management to install the AI Security Scanner thin client in one command. The thin client is a lightweight proxy that connects Claude Code to our cloud MCP server where all 27 security agents run.

⚠️API Key Needed

Complete Step 1 to get your API key, then return here for a personalized install command.

One-Command Installation

Run this command in your terminal with your API key from Step 1. The thin client connects to our cloud server where all 27 security agents run:

Claude MCP Add Command

claude mcp add ai-security-scanner -e AI_SECURITY_API_KEY=ciso_live_your_key_here -- uvx ai-security-mcp

📌 Don't Forget

Replace ciso_live_your_key_here with your actual API key from Step 1. Your API key will start with ciso_live_.

🔄 Important: Restart Required

After installing the MCP server, you must restart Claude Code for the new server to become active. Close and reopen Claude Code completely.

Verify Installation

After installation, verify the MCP server is added:

Verification Command

claude mcp list

Expected Output:

You should see ai-security-scanner in the list of installed MCP servers.

✓ ai-security-scanner
✓ Server connection verified

✨ What This Does

  • • Downloads and installs the ai-security-mcp thin client package via uvx
  • • Adds the MCP server to Claude Code automatically
  • • Sets up your API key as an environment variable
  • • Connects to our cloud MCP server at ai-security-mcp-fastmcp-production-amkk3ticjq-uc.a.run.app
  • • All 27 security agents run in the cloud - no local execution needed!
  • • No manual JSON configuration required!
3

Start Scanning

1 minute • Test your setup

Test your setup by running your first AI security scan in Claude Code.

Try this prompt in Claude Code:

Example Scan Prompt

Use the ai-security-agent MCP server to scan this repository for agentic AI security vulnerabilities. I want to see a comprehensive analysis covering all 17 OWASP ASI vulnerability types.

Expected Results:

  • • Executive summary with risk rating
  • • Analysis of all 17 OWASP ASI vulnerability types
  • • Specific findings with file locations
  • • Actionable remediation recommendations