1. Conduid
  2. Security
  3. Scanner
MCP server · Security

Scanner

Security scanner for MCP servers. Detects prompt injection, permission issues, and tool shadowing in Claude Desktop, Cursor, Windsurf, and other AI tools.

Unclaimed Apache-2.0 last commit 6 months ago llmmcpmcp-serversecurity
34Low

Scored 4 days ago · breakdown

About Scanner

Scanner is an MCP server published by Oabraham1 in the Security category: security scanner for MCP servers. Detects prompt injection, permission issues, and tool shadowing in Claude Desktop, Cursor, Windsurf, and other AI tools.

The repository has 5 stars and 0 forks, with the last commit 6 months ago. Six months or more without a commit doesn't mean the server is broken, but check the open issues before depending on it in production.

Install

From source
git clone https://github.com/Oabraham1/mcp-scanner

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Ask AI

✦

Ask AI about Scanner

Powered by Claude · Grounded in docs

I know everything about Scanner. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

Static checks derived from the repository 14 hours ago. Heuristics on the source, not an audit.

  • ✓README presentThe repository ships documentation.
  • ✓License declaredUsable under a stated license.
  • !Tests presentNo tests found. Behaviour changes may go unnoticed between versions.
  • ✓Dependencies pinnedA lockfile pins dependency versions.
  • ✓No dynamic code executionSource scan found no eval, dynamic Function, or shell execution.

1 check has not been run yet.

Releases

v0.1.1v0.1.1 · 7 Feb 2026Full Changelog: https://github.com/Oabraham1/mcp-scanner/compare/v0.1.0...v0.1.1
v0.1.0v0.1.0 · 7 Feb 2026Full Changelog: https://github.com/Oabraham1/mcp-scanner/commits/v0.1.0

README

mcp-scanner

CI Crates.io License Rust

Security scanner and proxy for MCP (Model Context Protocol) servers.

mcp-scanner discovers, scans, and proxies MCP servers configured across your AI tools (Claude Desktop, Cursor, Windsurf, VS Code, and more), detecting security vulnerabilities like prompt injection in tool descriptions, overly broad permissions, and suspicious changes.

Features

  • Auto-discovery: Finds MCP servers configured in Claude Desktop, Cursor, Windsurf, Zed, Cline, Continue, VS Code, Roo Code, and Claude Code
  • Security scanning: Detects prompt injection, permission scope issues, missing auth, tool shadowing, and description drift
  • STDIO proxy: Intercepts tool calls between clients and servers with rule-based filtering
  • Web dashboard: htmx-powered UI for viewing scan results and managing proxy rules
  • Audit logging: SQLite-backed logging of all proxied tool calls

Installation

Homebrew (macOS/Linux)

brew install oabraham1/tap/mcp-scanner

Shell Installer

curl -fsSL https://raw.githubusercontent.com/oabraham1/mcp-scanner/main/install.sh | sh

Download Binary

Download pre-built binaries from GitHub Releases.

Cargo (requires Rust)

cargo install mcp-scanner

Build from Source

git clone https://github.com/oabraham1/mcp-scanner
cd mcp-scanner
cargo build --release

Quick Start

# Scan all discovered MCP servers
mcp-scanner scan

# List discovered servers
mcp-scanner list

# Start the web dashboard
mcp-scanner serve

# Proxy a specific server
mcp-scanner proxy --server "npx -y @modelcontextprotocol/server-filesystem /"

CLI Reference

mcp-scanner scan

Scan MCP servers for security vulnerabilities.

mcp-scanner scan                           # Scan all discovered servers
mcp-scanner scan --client claude           # Scan only Claude Desktop servers
mcp-scanner scan --server "npx server.js"  # Scan a specific server command
mcp-scanner scan --config ./mcp.json       # Scan servers from config file
mcp-scanner scan --output json             # Output as JSON
mcp-scanner scan --output sarif            # Output as SARIF (for CI integration)

mcp-scanner list

List discovered MCP servers.

mcp-scanner list                    # List all servers
mcp-scanner list --client cursor    # List only Cursor servers

mcp-scanner serve

Start the web dashboard and API server.

mcp-scanner serve                   # Start on localhost:9191
mcp-scanner serve --port 8080       # Use custom port
mcp-scanner serve --headless        # Don't open browser

mcp-scanner proxy

Proxy an MCP server with filtering and audit logging.

mcp-scanner proxy --server "npx -y @modelcontextprotocol/server-filesystem /"

To use the proxy, update your client config to point to mcp-scanner:

{
  "mcpServers": {
    "filesystem": {
      "command": "mcp-scanner",
      "args": ["proxy", "--server", "npx -y @modelcontextprotocol/server-filesystem /"]
    }
  }
}

mcp-scanner init

Create default configuration.

mcp-scanner init           # Create ~/.mcp-scanner/config.toml
mcp-scanner init --force   # Overwrite existing config

mcp-scanner completions

Generate shell completions.

mcp-scanner completions --shell bash >> ~/.bashrc
mcp-scanner completions --shell zsh >> ~/.zshrc
mcp-scanner completions --shell fish >> ~/.config/fish/completions/mcp-scanner.fish

Threat Categories

mcp-scanner detects the following security issues:

Description Injection (Critical/High)

Prompt injection patterns in tool descriptions, including:

  • "Ignore previous instructions" patterns
  • Hidden Unicode characters
  • Base64-encoded payloads
  • System prompt injection attempts

Permission Scope (High/Medium)

Overly broad capabilities:

  • Arbitrary code execution
  • Root filesystem access
  • Unrestricted network access
  • Database query access

No Auth (Critical for remote, Info for local)

Servers without authentication:

  • Remote servers without auth tokens (Critical)
  • Local servers without env-based auth (Info)

Tool Shadowing (High/Medium)

Name conflicts across servers:

  • Exact name collisions
  • Similar names (potential typosquatting)

Description Drift (High/Medium)

Changes since last scan:

  • Modified tool descriptions
  • Added/removed tools

Configuration

Config file location: ~/.mcp-scanner/config.toml

[scan]
timeout = 30  # seconds per server

[output]
format = "table"  # table, json, sarif

API Endpoints

The web server exposes a JSON API:

  • GET /api/health - Health check
  • GET /api/servers - List discovered servers
  • POST /api/scan - Run a scan
  • GET /api/audit - List audit log entries
  • GET /api/rules - List proxy rules
  • POST /api/rules - Create proxy rule
  • PUT /api/rules/:id - Update proxy rule
  • DELETE /api/rules/:id - Delete proxy rule

Data Storage

mcp-scanner stores data in ~/.mcp-scanner/:

  • mcp-scanner.db - SQLite database (audit logs, scan results, rules)
  • snapshots/ - Tool description snapshots for drift detection
  • config.toml - Configuration file

Supported Clients

Client Config Path
Claude Desktop ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)
Cursor ~/.cursor/mcp.json
Windsurf ~/.codeium/windsurf/mcp_config.json
Zed ~/.config/zed/settings.json
Cline ~/.config/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json
Continue ~/.continue/config.json
VS Code .vscode/mcp.json
Roo Code ~/.config/Code/User/globalStorage/rooveterinaryinc.roo-cline/settings/mcp_settings.json
Claude Code ~/.claude/settings.json or .mcp.json

License

Licensed under either of:

at your option.

README fetched from the repository 14 hours ago. The original is authoritative.

Questions

About Scanner

How do I install Scanner?

No package has been published for this server, so install it from source: clone the repository and follow its README.

Is Scanner safe to use with an AI agent?

Its trust score is 34 out of 100 (low). It passes 4 of the 5 static security checks run so far; the failures are listed above. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Scanner still maintained?

The last commit was 6 months ago and 15 open issues. That's long enough that you should check whether the maintainer is responding to issues before depending on it.