Argus MCP Server

How to use

You interact with Argus through an MCP client to analyze repositories, detect languages, run static checks, and perform security scans. Start by pointing Argus at a repository, choose the analysis you want, and review structured reports that cover code quality, language-specific findings, and security issues. You can run analyses on multiple branches, compare changes, and generate comprehensive results that guide remediation.

How to install

Prerequisites you need before installing Argus include Python 3.8 or newer, Git, and the libmagic library. You also need a functioning target operating system (macOS or Linux) to install system dependencies.

# macOS
brew install libmagic

# Linux (Ubuntu/Debian)
sudo apt-get update
sudo apt-get install -y libmagic1

Install Argus using the MCP runtime. Run the following command to install and start Argus through the MCP runtime interface.

uvx argus

Optional: if you need to run Argus from a specific configuration, you can use the MCP command channel described in the configuration examples. Ensure you have internet access and the MCP runtime installed on your system.

Configuration and runtime details

Argus provides a configuration snippet that defines how the server is started, which operations are allowed, and how long the server can run before timing out. Use the following JSON snippet to configure the MCP client or orchestrator to initialize Argus with the appropriate runtime parameters.

{
  "command": "uvx",
  "args": [
    "--from",
    "git+https://github.com/athapong/argus",
    "argus"
  ],
  "alwaysAllow": [
    "get_commit_history",
    "enumerate_branches",
    "compare_git_changes",
    "analyze_code_quality",
    "security_scan_repository"  
  ],
  "timeout": 300
}

Environment and usage notes

You can control system checks and PATH management through environment variables. For example, set SKIP_SYSTEM_CHECK to bypass system dependency checks during initial setup. Argus will also automatically adjust PATH to include newly installed toolchain binaries.

Tools and capabilities

Argus supports a range of language-specific analysis tools and a security scanner. Core capabilities include multi-language support, branch-aware analysis, commit history insights, repository structure visualization, and comprehensive security reports.

Troubleshooting and tips

If you encounter dependency installation issues, Argus will emit warnings and attempt graceful fallbacks. Language detection will fall back to the explicitly provided language if automatic detection fails, and tool execution errors will be reported in the results structure.

gocyclo

Go language complexity analysis used alongside golangci-lint in multi-language analysis

golangci-lint

Go static analysis tool for code quality and linting

PMD

Java static analysis to identify potential problems and code quality issues

Pylint

Python linting tool for code quality and style checks

Bandit

Python security scanner focusing on known vulnerabilities and risky patterns

ESLint

JavaScript/TypeScript linting tool to enforce code quality and consistency

Trivy

Security scanner for vulnerability detection in container images and files

analyse_repository_structure

Function to analyze repository structure and generate insights

analyze_code_quality

Function to perform language-appropriate quality checks across supported languages

security_scan_repository

Function to run security scans on a repository and produce reports

get_commit_history

Retrieve commit history for a repository

enumerate_branches

List branches within the repository

compare_git_changes

Compare changes between two branches or commits