SmuggleCheck

Overview

SmuggleCheck is a Model Context Protocol server that detects hidden prompt-injection text across multiple obfuscation techniques including ASCII smuggling, zero-width and emoji steganography, bidirectional text attacks, and homoglyph substitution. It integrates directly into AI assistants as callable tools to help identify malicious or concealed inputs.

SmuggleCheck runs as a remote hosted service at https://smugglecheck.mike-tusa.workers.dev/mcp, requiring only an API key for authentication. The latest release is v0.1.2. No local installation or infrastructure setup is needed — simply point your MCP client to the endpoint and connect.

SmuggleCheck is listed in the AI & Agents category and works with Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, Zed, and other MCP clients. If you maintain SmuggleCheck, claim this listing to verify ownership and keep information up to date.

Installation

Settings → Connectors → Add custom connector
Name: SmuggleCheck
URL:  https://smugglecheck.mike-tusa.workers.dev/mcp