n0s1-mcp

An MCP server that exposes n0s1 secret-scanning capabilities as tools for AI assistants (Claude, Cursor, etc.).

Scan Jira, Confluence, Slack, GitHub, GitLab, Zendesk, Linear, Asana, Wrike, and local filesystems for leaked secrets — directly from your AI workflow.

Quickstart

No install required. Add this to your MCP client config and run via uvx:

{
  "mcpServers": {
    "n0s1": {
      "command": "uvx",
      "args": ["n0s1-mcp"]
    }
  }
}

For Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json
For Claude Code: .claude/mcp.json in your project, or ~/.claude/mcp.json globally.

Available Tools

All scan_* tools accept these optional parameters:

Pass wait_minutes to analyze_report (or directly on a scan tool alongside ai_analysis) to block until analysis completes.

Environment Variables

Credentials can be passed as tool arguments or pre-set as environment variables:

Example with env vars pre-configured:

{
  "mcpServers": {
    "n0s1": {
      "command": "uvx",
      "args": ["n0s1-mcp"],
      "env": {
        "GITHUB_TOKEN": "ghp_...",
        "JIRA_TOKEN": "ATATT..."
      }
    }
  }
}

Usage Examples

Once connected, ask your AI assistant:

• "Scan my Jira project SEC for leaked secrets"
• "Check the GitHub org mycompany for exposed API keys"
• "Scan the /home/user/project directory for secrets"
• "Run an AI analysis on the scan report abc123"

For full parameter reference and AI analysis workflow details, see docs/ai.md.

Publishing to PyPI

pip install hatch
hatch build
hatch publish

License

GNU General Public License v3 — same as n0s1.