MEDUSA CLI
A security scanner CLI for code and AI agents: vets .claude, hooks, MCP configs and SKILL.md before cloning and finds leaked keys
- 1,018 ★ on the repository
- Last commit
- AGPL-3.0
Extended access
Why this level
- Runs code and reads project files and histories
- Reaches the network to vet remote repos
- In-place redaction changes files
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/medusaDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs code and reads project files and histories
- Reaches the network to vet remote repos
- In-place redaction changes files
This is third-party code. Review the repository files before installing.
What it does
MEDUSA scans projects and repositories against a large rule set aimed at vulnerabilities in code, AI apps, agents and MCP servers. With medusa scan --git you can vet any GitHub repo before cloning: the scanner recognizes dozens of AI editor config files that become attack vectors, including .cursor/mcp.json, CLAUDE.md, SKILL.md and AGENTS.md. medusa secrets scan finds leaked keys and tokens in Claude Code, Cursor and Copilot chat histories and in shell history, with interactive in-place redaction. It works right after install with no required external tools, and auto-detects installed linters such as bandit and eslint.
Who it is for. For developers and security engineers who install third-party skills, MCP servers and repos and want to vet them first.
Good fit when
- Before cloning a third-party repo with skills or MCP
- You need to check .claude, hooks and editor configs for anything suspicious
- You need to find leaked keys in chat and shell history
Not a fit when
- You need a full manual security audit
- The project is outside supported languages and formats
Example request
Use medusa to vet this GitHub repo before I clone itLimitations
It is a static signature and rule scanner: it catches known patterns but does not replace manual review or guarantee a project is safe. In-place secret redaction changes your files, so review the edits. Model scanning and external linters need extra tools.
How to disable. Remove the package: pip uninstall medusa-security, or delete the virtual environment that holds it.
FAQ
Can I vet a repo without cloning it?
Yes, medusa scan --git owner/repo checks a remote repo and recognizes dangerous AI editor configs.
Does it change my files?
A normal scan only reads. Redacting leaked secrets in place changes files and is confirmed by you.
Discussion
No questions or reviews yet.
Related
CLI toolsA scanner that finds leaked keys, dangerous access and connected models in MCP configs of Claude Desktop, Cursor and VS Code.
A local scanner flagging malicious packages and risky permissions in dependencies, agent instructions and MCP configs.
AgentShield audits the .claude folder for plaintext keys, broad permissions, dangerous hooks and risky MCP servers and grades the setup from A to F
A local dashboard for Claude Code, Codex CLI and OpenCode: what the agent loads, its token cost and an MCP server security scan
Security scanner for coding agents: checks code, MCP servers, prompts and model-invented packages, and gives an A to F grade.
A security scanner for AI-written code: secrets, dependencies, MCP and CI, findings investigated with evidence, and fixes you approve
README in short
The README presents MEDUSA as an AI-first security scanner for agents and MCP that works right after pip install. It describes scanning a project, a remote repo and finding leaked secrets in chat and shell histories. It lists categories of dangerous editor config files and known supply chain attacks. There are extra modes for ML models and external linters. AGPL-3.0 licensed.