About MCP Security Analyst
MCP Security Analyst provides code and dependency security checks for developers and AI assistants. It combines OSV.dev vulnerability lookups, Go AST analysis, pattern checks, and Gitleaks secret detection.
Key features
- 01Check package versions against OSV.dev
- 02Analyze Go code with AST checks
- 03Detect secrets with Gitleaks
- 04Identify vulnerability patterns
- 05Redact detected secrets in output
Use cases
- 01Check dependencies in go.mod for known issues
- 02Scan a repository for exposed credentials
- 03Review code for injection and deserialization risks
- 04Combine code, secret, and dependency checks
MCP support
The MCP server exposes check_vulnerabilities, analyze_security, and scan_secrets tools for vulnerability and code security analysis.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEMCP Security Analyst
A Model Context Protocol (MCP) server providing comprehensive security analysis capabilities through integration with OSV.dev vulnerability database and native Go-based code analysis and secret detection engines.
Features
- Supply Chain Vulnerability Analysis: Integration with OSV.dev API for dependency vulnerability assessment
- Secret Detection: Gitleaks v8 integration with 100+ built-in detection rules for credentials and API keys
- Static Code Analysis: AST-based Go code analysis for security anti-patterns
- Pattern Matching: Regex-based detection for common security vulnerabilities
- MCP Protocol Support: Standard protocol implementation for AI assistant integration
- Community-Vetted Rules: Gitleaks patterns maintained by the security community
Requirements
Core Requirements
make deps
make install
Build Dependencies
- Go 1.25.4 or later
- github.com/mark3labs/mcp-go
- github.com/zricethezav/gitleaks/v8
Installation
make deps # Install Go module dependencies
make build # Compile binary
make install # Install to /usr/local/bin
make run # Build and execute server
make clean # Remove build artifacts
The mcp-osv binary communicates via stdin/stdout using the MCP protocol.
IDE Configuration
Cursor IDE
Navigate to Configuration > MCP and add:
{
"mcpServers": {
"security_analyst": {
"name": "Security Analyst",
"type": "stdio",
"command": "/usr/local/bin/mcp-osv"
}
}
}
Claude Desktop
Edit the MCP configuration file at Settings > Developer:
{
"mcpServers": {
"mcp_osv": {
"command": "/usr/local/bin/mcp-osv",
"args": []
}
}
}
Available Tools
The server exposes three MCP tools for security analysis:
check_vulnerabilities
Query OSV.dev database for known vulnerabilities in specific package versions.
Parameters:
package_name(string, required): Package identifierversion(string, required): Version string
Functionality:
- Rate-limited API requests (1 request/second)
- HTTP timeout protection (10 seconds)
- JSON response parsing
- Vulnerability detail extraction
analyze_security
Comprehensive security analysis combining multiple detection engines.
Parameters:
file_path(string, required): Target file or directory path
Analysis Components:
- Native Go AST-based code analysis
- Gitleaks v8 secret detection with 100+ rules
- OSV.dev vulnerability checks for dependencies (go.mod files)
- Pattern-based vulnerability detection
Detected Issues:
- Command injection vectors
- Deserialization vulnerabilities
- SQL injection patterns
- Hardcoded credentials
- API keys and tokens
- Private keys and certificates
- Database connection strings
scan_secrets
Dedicated secret detection using Gitleaks v8 with 100+ community-maintained detection rules.
Parameters:
path(string, required): Target file, directory, or repository pathscan_git_history(boolean, optional): Enable git history scanning (default: false)
Detection Capabilities (100+ patterns):
- AWS Access Keys, Secret Keys, Session Tokens
- GitHub Personal Access Tokens, OAuth tokens
- Google Cloud Platform API keys
- Azure credentials and connection strings
- Slack tokens and webhooks
- Stripe API keys
- Private SSH/PGP/RSA keys
- JWT tokens
- Database connection strings (PostgreSQL, MySQL, MongoDB)
- Generic API keys with entropy analysis
- And 90+ more patterns maintained by the security community
Output: Partial secret redaction for secure display (first 4 + last 4 characters)
Integration Patterns
The MCP server enables AI assistants to perform security analysis through natural language requests:
Dependency Vulnerability Scanning:
Request: "Check dependencies in go.mod for vulnerabilities"
Tool Execution: analyze_security -> OSV.dev API queries
Response: Vulnerability report with CVE details
Secret Detection:
Request: "Scan repository for exposed credentials"
Tool Execution: scan_secrets -> Pattern matching + entropy analysis
Response: Detected secrets with file locations and types
Comprehensive Audit:
Request: "Perform full security analysis"
Tool Execution: analyze_security -> All detection engines
Response: Combined report (code issues + secrets + vulnerabilities)
Security Implementation Details
Rate Limiting
OSV.dev API requests are rate-limited at 1 request per second using golang.org/x/time/rate limiter to prevent service throttling.
Input Validation
All file paths undergo sanitization to prevent directory traversal attacks:
- Path cleaning via filepath.Clean()
- Directory traversal pattern detection
- Existence verification
Secret Redaction
Detected secrets are partially redacted before display:
- Secrets <= 8 characters: Full redaction
- Secrets > 8 characters: First 4 + "***" + Last 4 characters
Gitleaks Integration
Secret detection powered by Gitleaks v8:
- 100+ community-maintained detection rules
- Entropy analysis for high-randomness strings
- Keyword-based pre-filtering for performance
- Regular updates for new secret types
Adding Security Rules
To extend detection capabilities:
- Secrets and credential detection: Gitleaks rules are maintained upstream at gitleaks/gitleaks
- Code Analysis: Extend AST inspection in
runGoCodeAnalysis() - Pattern Matching: Regex patterns can be added to
analyzeFile()checks map, create a branch and PR explaining them to get merged
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.