What it does
This MCP server connects VibeScan to Claude Code. It exposes two tools: vibescan_scan, which scans a project for secrets, dangerous code patterns and git hygiene problems, and vibescan_rules, which lists the 17 detection rules. The secret checks cover 14 categories, including env files, cloud credentials, Docker and infrastructure files, CI/CD pipelines, SSH keys and mobile files. Code checks flag patterns such as eval(), exec(), shell injection, SQL injection, pickle and innerHTML. Git checks flag a missing .gitignore and unignored .env, .pem or .key files.
It suits a developer who wants to check a project for leaked credentials or common security issues from inside Claude Code. The README gives the example request "check for leaked secrets" as a way to start a scan in plain language.
The server is added to a project's .mcp.json file, after which Claude can run the scan on request. According to the README, all scanning runs locally, so the code is not sent off the machine. The project is released under the MIT license.
Limitations and requirements
Set up for use with Claude Code, with the connection configured in a .mcp.json file. No other requirements are stated.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.Aguantar/vibescan-mcp-server
MCP server for VibeScan — scan projects for leaked secrets and security issues
v0.1.0As shown on registry.modelcontextprotocol.io when we last read it