Veridano/veridano-mcp-server

Veridano Intelligence API

Government cybersecurity intelligence MCP server for AI agents

Access 14 US, UK, and EU government cybersecurity agencies through a hosted MCP server with 6,000+ curated intelligence documents including CISA KEV, MITRE ATT&CK, NIST SP 800, and international threat intelligence.

⚡ Quick Start

For Claude Desktop

Step 1: Open Claude Desktop application

Step 2: Click Settings → Features → Model Context Protocol

Step 3: Click Add custom connector

Step 4: Fill in the connector details:

• Name: Veridano
• Remote MCP server URL: https://7lqg8v66p1.execute-api.us-east-1.amazonaws.com/prod/mcp

Step 5: Click Add then restart Claude Desktop

That's it! No downloads required - Claude connects directly to the hosted MCP server.

For ChatGPT

Step 1: Go to ChatGPT Settings > Beta Features > Model Context Protocol

Step 2: Add Custom MCP Server:

• Name: Veridano Intelligence
• Server URL: https://7lqg8v66p1.execute-api.us-east-1.amazonaws.com/prod/mcp

Step 3: Enable the server

That's it! Start querying: "Use veridano_search to find CISA ransomware advisories"

🔗 Direct API Access

For direct REST API access without MCP:

curl -X POST "https://7lqg8v66p1.execute-api.us-east-1.amazonaws.com/prod/api/search" \
  -H "Content-Type: application/json" \
  -d '{"query": "CISA ransomware advisory", "top_k": 5}'

🎯 Overview

Veridano provides AI agents with comprehensive access to U.S. government cybersecurity intelligence through a unified MCP server interface.

Key Capabilities

• Real-time intelligence search across 14 US/UK/EU government agencies
• 7 MCP tools for threat intelligence, CVE analysis, and APT tracking
• Complete CISA KEV catalog - All known exploited vulnerabilities with BOD 22-01 compliance
• MITRE ATT&CK integration - 1,450+ techniques, groups, and mitigations
• International coverage - NCSC (UK) and ENISA (EU) threat intelligence
• Vector similarity search with semantic understanding via AWS Titan V2 embeddings
• Query expansion - Automatic abbreviation expansion (BOD, KEV, APT, etc.)
• Citation support - Proper APA/MLA citations for government sources
• Zero-config setup - Works immediately after MCP server connection
• High-performance - Sub-second response times with Aurora + pgvector

🏛️ Data Sources (14 Agencies)

The platform continuously monitors and indexes content from:

US Government Sources

Frameworks & International

Total Coverage: ~6,000+ documents across 14 agencies

🔧 MCP Tools Available

Once connected, AI agents have access to 7 specialized tools:

1. veridano_search - Semantic search across all sources

   • Parameters: query, top_k, sources
   • Best for: General threat intelligence queries

2. get_latest - Recent documents (last N hours)

   • Parameters: hours, limit, sources, critical_only
   • Best for: "What's new?", breaking alerts

3. batch_search - Multiple queries at once

   • Parameters: queries (array of query objects)
   • Best for: Multi-topic research, 3-10x faster

4. analyze_threat_trends - Historical trend analysis

   • Parameters: threat_category, days
   • Best for: Temporal patterns, activity analysis

5. compare_vulnerabilities - Side-by-side CVE comparison

   • Parameters: cve_ids (array)
   • Best for: Vulnerability prioritization

6. get_remediation_guidance - Mitigation steps

   • Parameters: threat_or_cve
   • Best for: Actionable remediation advice

7. track_threat_actor - APT group tracking

   • Parameters: actor_name
   • Best for: APT profiling and attribution

📊 Search Parameters

veridano_search parameters:

• query (string) - Search query (supports query expansion)
• top_k (integer) - Number of results (default: 5)
• min_score (float) - Similarity score 0.0-1.0 (default: 0.6)
• sources (array) - Filter: ["CISA", "FBI", "NIST", "DHS", "NSA", "USCYBERCOM", "MITRE", "NCSC", "ENISA", "Treasury", "DoD", "White House", "US-CERT", "ICS-CERT"]

📝 Example Queries

Vulnerability Research:

• "Use veridano_search to find CVE-2024-3400 across all sources"
• "Compare CVE-2024-3400 and CVE-2023-44487 using compare_vulnerabilities"
• "Get remediation guidance for Log4Shell"
• "Find BOD 22-01 compliance requirements from CISA"

Threat Intelligence:

• "Track APT29 activity using track_threat_actor"
• "Use veridano_search for ransomware guidance from CISA and FBI"
• "Analyze ransomware trends over the last 90 days"
• "Show me critical alerts from the last 24 hours using get_latest"

Compliance & Frameworks:

• "Search for NIST SP 800-207 zero trust architecture"
• "Find all CISA binding operational directives"
• "Search for MITRE ATT&CK techniques used by APT29"
• "Get recent UK threat intelligence from NCSC"

🏗️ Architecture

AI Agent → MCP Client → Veridano MCP Server → Enhanced Scraper Network → Government Data Sources

Data Collection (September 2025)

Veridano operates 20 specialized scrapers for comprehensive government intelligence:

Enhanced Scrapers (8):

• CISA RSS (4 hours), CISA KEV (6 hours), CISA BOD (daily)
• FBI IC3 (12 hours), NIST Comprehensive (6 hours), NIST SP 800 (daily)
• DoD/DC3 (daily), Treasury/FinCEN (daily)

Specialized Intelligence (3):

• MITRE ATT&CK (weekly) - 1,450+ techniques, groups, mitigations
• International Sources (daily) - NCSC (UK), ENISA (EU)
• Real-time Alerting - Critical event notifications

Legacy Scrapers (9):

• NSA, USCYBERCOM, White House, US-CERT, ICS-CERT, NVD, FedRAMP

Technical Stack

• Compute: AWS Lambda (29 functions, Python 3.11, ARM64)
• AI/ML: Amazon Bedrock (Titan V2 Embeddings, 1024 dimensions)
• Database: Aurora Serverless v2 with pgvector for semantic search
• Connection Pooling: RDS Proxy (40-60% overhead reduction)
• Storage: S3 (10 buckets, lifecycle policies)
• Protocol: Model Context Protocol (MCP) + REST APIs
• Scheduling: EventBridge (14 schedules, 1-hour to weekly)
• Monitoring: CloudWatch Dashboard, SNS Alerting
• Quality: Data validation pipeline with duplicate detection

🚀 Usage Examples

Once configured, you can immediately start querying cybersecurity intelligence:

Ask Claude Desktop:

• "Use veridano_search to find CISA ransomware guidance"
• "Show me critical cybersecurity alerts from the last 24 hours using get_latest"
• "Track APT29 activity across government reports"
• "Compare CVE-2024-3400 with CVE-2023-44487 using compare_vulnerabilities"
• "Get remediation guidance for Log4Shell"
• "Analyze ransomware trends over the last 90 days"
• "Search for MITRE ATT&CK techniques used by APT29"
• "Find UK threat intelligence from NCSC about state-sponsored attacks"

📊 Performance

• Response time: <1 second average (150-300ms typical)
• Database: 6,000+ curated government cybersecurity documents
• Coverage: 14 agencies (US, UK, EU)
• MITRE ATT&CK: Complete (1,450+ documents)
• Updates: Every 1-6 hours for real-time sources, daily/weekly for comprehensive sources
• Rate Limit: 1,000 requests/second
• Uptime: 99.9% target
• Tools: 7 MCP tools + 5 analytics APIs
• Infrastructure: 29 Lambda functions, RDS Proxy, Aurora Serverless v2

📝 Response Format

{
  "results": [
    {
      "document_id": "cisa-aa24-131",
      "title": "CISA Ransomware Guide",
      "quick_summary": "First 500 characters for rapid assessment...",
      "content": "Full content...",
      "source": "CISA",
      "url": "https://...",
      "published_date": "2024-05-10",
      "relevance_score": 0.94,
      "citation": {
        "apa": "CISA. (2024). Ransomware Guide 2024. Retrieved from https://...",
        "mla": "CISA. \"Ransomware Guide 2024.\" 10 May 2024, https://...",
        "permalink": "https://www.cisa.gov/..."
      },
      "source_metadata": {
        "organization": "CISA",
        "authority_level": "federal_mandate",
        "reliability_score": 1.0,
        "document_type": "cybersecurity_advisory",
        "classification": "public"
      }
    }
  ],
  "search_metadata": {
    "sources_searched": ["CISA", "FBI"],
    "embedding_model": "AWS_Titan_V2",
    "min_score_threshold": 0.6,
    "search_backend": "Aurora_pgvector"
  }
}

⚡ Usage Guidelines

• Concurrent sessions: 1000+ supported
• Query complexity: No limits
• Result size: Up to 50 documents per query
• Fair usage: Please cache results when appropriate

💡 Alternative Setup Methods

Local Development Setup

For developers who want to modify or test the MCP client locally:

git clone https://github.com/Veridano/veridano-mcp-server.git
cd veridano-mcp-server

# For Claude Desktop development (legacy method):
{
  "mcpServers": {
    "veridano-local": {
      "command": "python",
      "args": ["/full/path/to/veridano-mcp-server/mcp_client.py"]
    }
  }
}

Note: Custom Connector (direct URL) is recommended for regular usage.

Enterprise Options

Contact enterprise@veridano.com for:

• Custom authentication requirements
• Private deployment options
• Volume licensing and SLAs

📚 Documentation

• - Complete MCP tool documentation
• - Detailed installation instructions
• - Sample agent implementations
• - Optimization best practices
• - Common issues and solutions

🤝 Support

• Issues: GitHub Issues
• Enterprise Support: enterprise@veridano.com

📄 License

MIT License - See LICENSE file for details.

Built for AI Agents. Powered by U.S. Government Intelligence.