Safeguard MCP Server

The Safeguard MCP Server enables AI assistants like Claude Desktop, ChatGPT, and Grok to interact with Safeguard's enterprise-grade vulnerability management and SBOM analysis platform using natural language.

Overview

Model Context Protocol (MCP) is an open standard that allows AI assistants to securely connect to external services. The Safeguard MCP Server exposes 35+ tools for vulnerability scanning, SBOM analysis, AI-powered remediation, SCM integration, and policy enforcement.

Key Features

Quick Start

1. Get API Key

Sign up for a Safeguard account and generate an API key:

1. Visit app.safeguard.sh/signup
2. Navigate to Settings → API Keys
3. Click Generate API Key
4. Copy your API key (starts with sg_api_)

2. Configure Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "safeguard": {
      "url": "https://mcp.safeguard.sh/mcp/sse",
      "transport": "sse",
      "env": {
        "SAFEGUARD_API_KEY": "sg_api_your_key_here"
      }
    }
  }
}

Config file locations:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

3. Restart Claude Desktop

Restart Claude Desktop to load the MCP server connection.

4. Start Using

Try these prompts in Claude Desktop:

"Find all critical vulnerabilities in my projects"
"Generate an SBOM for my latest deployment"
"Show me packages with high EPSS scores"
"Remediate vulnerabilities in my package.json"
"What's my current security risk score?"

Available Tools

The Safeguard MCP Server provides 35+ tools across these categories:

Vulnerability Management

• safeguard_find_vulnerabilities - Search for CVEs and vulnerabilities
• safeguard_get_vulnerability - Get detailed CVE information
• safeguard_check_exploitability - Check if CVE has known exploits (EPSS)
• safeguard_get_remediation - Get fix recommendations

SBOM Operations

• safeguard_list_sboms - List all SBOMs in your account
• safeguard_get_sbom - Retrieve SBOM details
• safeguard_create_sbom - Generate new SBOM
• safeguard_compare_sboms - Compare two SBOMs for differences
• safeguard_analyze_sbom - Get security analysis of SBOM
• safeguard_export_sbom - Export in CycloneDX/SPDX formats

Package Analysis

• safeguard_find_packages - Search packages across all SBOMs
• safeguard_get_package - Get package details and metadata
• safeguard_get_package_vulnerabilities - List vulnerabilities in package
• safeguard_get_package_licenses - Check package licenses

Griffin AI Remediation

• safeguard_remediate_npm - Fix npm package.json vulnerabilities
• safeguard_remediate_pip - Fix Python requirements.txt vulnerabilities
• safeguard_remediate_maven - Fix Maven pom.xml vulnerabilities
• safeguard_remediate_go - Fix Go go.mod vulnerabilities
• safeguard_remediate_cargo - Fix Rust Cargo.toml vulnerabilities
• safeguard_get_remediation_plan - Get AI-generated remediation strategy

Project Management

• safeguard_list_projects - List all projects
• safeguard_get_project - Get project details
• safeguard_create_project - Create new project
• safeguard_get_project_sboms - Get all SBOMs for a project
• safeguard_get_project_metrics - Get security metrics

Policy & Compliance

• safeguard_list_gates - List security policy gates
• safeguard_evaluate_gate - Check if SBOM passes policy
• safeguard_create_gate - Create new policy gate
• safeguard_get_compliance_report - Generate compliance report

SCM Integration

• safeguard_add_github_integration - Connect GitHub repository
• safeguard_add_gitlab_integration - Connect GitLab repository
• safeguard_add_bitbucket_integration - Connect Bitbucket repository
• safeguard_add_azuredevops_integration - Connect Azure DevOps
• safeguard_scan_repository - Scan SCM repository for vulnerabilities
• safeguard_list_scm_projects - List integrated repositories

Natural Language Query

• safeguard_ask - Ask questions in natural language
• safeguard_chat - Multi-turn conversation with Griffin AI

Example Use Cases

1. Security Review Workflow

User: "Review the security of my production SBOMs"

Claude uses:
1. safeguard_list_sboms (filter: production)
2. safeguard_analyze_sbom (for each SBOM)
3. safeguard_find_vulnerabilities (critical severity)
4. Summarizes findings with risk assessment

2. Vulnerability Remediation

User: "Fix the vulnerabilities in my package.json"

Claude uses:
1. safeguard_get_project (current project)
2. safeguard_find_vulnerabilities (in project)
3. safeguard_remediate_npm (with package.json content)
4. Returns updated package.json with fixes

3. Compliance Check

User: "Check if my latest SBOM meets FedRAMP requirements"

Claude uses:
1. safeguard_list_gates (type: compliance)
2. safeguard_evaluate_gate (FedRAMP policy)
3. safeguard_get_compliance_report
4. Lists passing/failing criteria

4. Repository Scanning

User: "Scan my GitHub repository for security issues"

Claude uses:
1. safeguard_add_github_integration
2. safeguard_scan_repository
3. safeguard_find_vulnerabilities (in scan results)
4. safeguard_get_remediation_plan

Authentication

The MCP server supports API key authentication. Your API key is securely transmitted with each request and validated against Safeguard's Auth Service.

Security Best Practices

1. Never commit API keys to version control
2. Use environment variables in Claude Desktop config
3. Rotate keys regularly (every 90 days)
4. Use separate keys for different environments (dev/staging/prod)
5. Revoke compromised keys immediately in Settings

Multi-Product Access

The MCP server provides access to all Safeguard products:

Data Handling

The Safeguard MCP Server:

✅ Only accesses data explicitly requested by the user through prompts ✅ Encrypts all data transmission using HTTPS/TLS ✅ Is stateless - the MCP server itself stores no data ✅ Is GDPR compliant with data retention per terms of service

Pricing

Visit safeguard.sh/pricing for details.

Troubleshooting

Connection Issues

Problem: "Failed to connect to MCP server"

Solutions:

1. Verify API key is correct and not expired
2. Check internet connection
3. Ensure Claude Desktop is up to date
4. Check MCP server status at status.safeguard.sh

Authentication Errors

Problem: "Unauthorized" or "Invalid API key"

Solutions:

1. Regenerate API key in Settings → API Keys
2. Update claude_desktop_config.json with new key
3. Restart Claude Desktop
4. Verify key starts with sg_api_

Tool Errors

Problem: Tool returns an error or unexpected result

Solutions:

1. Check tool input parameters match expected format
2. Verify you have permissions for the requested operation
3. Check API rate limits (100/day for free tier)
4. Contact support at hi@safeguard.sh

Rate Limiting

Free tier users are limited to 100 API calls per day. Upgrade to Pro for unlimited access.

Support

• Email: hi@safeguard.sh
• Documentation: docs.safeguard.sh
• Status Page: status.safeguard.sh
• Response Time: 24 hours (free), 4 hours (paid)

System Requirements

Technical Details

• Endpoint: https://mcp.safeguard.sh/mcp/sse
• Transport: Server-Sent Events (SSE)
• Protocol: HTTPS only
• Uptime: 99.9% SLA (Enterprise)
• Response Time: < 2 seconds average
• Data Location: United States (AWS)

Privacy & Compliance

• Terms of Service: safeguard.sh/company/terms-and-conditions
• Privacy Policy: safeguard.sh/company/privacy-policies
• Compliance: SOC 2 Type II, GDPR, CCPA, ISO 27001

Related Developer Tools

IDE Extensions

Get real-time vulnerability detection directly in your IDE:

• VS Code Extension - Catch vulnerabilities as you write code with instant feedback and one-click remediation
• Cursor Extension - Same extension available for Cursor IDE via Open VSX
• JetBrains Plugin - For all JetBrains IDEs: IntelliJ IDEA, PyCharm, WebStorm, PhpStorm, GoLand, RubyMine, CLion, Rider, DataGrip, AppCode, and Android Studio

Next Steps

1. Sign up for Safeguard
2. Generate your API key
3. Configure Claude Desktop (see Quick Start above)
4. Try the example prompts
5. Explore the full API Reference

Developer Resources

Official repositories and packages:

• MCP Server: mcp.safeguard.sh - Model Context Protocol server for AI assistants
• NPM: @safeguard-sh - Official NPM packages
• GitHub: Safeguard-sh - Open source tools and integrations

Feedback

We're constantly improving the MCP server. Share your feedback:

• Feature requests: hi@safeguard.sh
• Bug reports: hi@safeguard.sh
• General questions: hi@safeguard.sh

Ready to get started? Sign up now and connect Claude to your security workflow in under 5 minutes.