Cursor Integration

Use Inkog directly within Cursor IDE for security analysis while you code.

Installation

{
  "mcpServers": {
    "inkog": {
      "command": "npx",
      "args": ["-y", "@inkog-io/mcp"],
      "env": {
        "INKOG_API_KEY": "sk_live_your_api_key"
      }
    }
  }
}

Usage

With Inkog installed, you can ask Cursor’s AI to:

Scan the Current File

“Use Inkog to scan this file for security vulnerabilities”

Scan the Whole Project

“Scan this project for AI agent security issues”

Fix Vulnerabilities

“Explain and fix the prompt injection vulnerability on line 42”

Generate Compliance Reports

“Generate an EU AI Act compliance report for this agent”

Example Conversation

You: Scan this CrewAI agent for security issues

Cursor: I'll use Inkog to scan your agent code.

[Uses inkog_scan tool]

Found 3 security issues:

1. [HIGH] Prompt Injection Risk (line 24)
   User input is passed directly to the system prompt

2. [MEDIUM] Missing Rate Limit (line 45)
   No rate limiting on LLM API calls

3. [LOW] Sensitive Data Logging (line 67)
   API key may be logged in debug output

Would you like me to help fix these issues?

Benefits

• Real-time Security: Get security feedback as you write code
• Contextual Fixes: Cursor understands your code structure
• One-Click Remediation: Apply fixes directly in the editor
• Governance Validation: Ensure code matches AGENTS.md

Troubleshooting

Next Steps

• MCP Server Reference - Full tool documentation
• Vulnerabilities - Learn about detected vulnerability types