# PromptScan MCP server

Prompt injection detection API for AI agents. Scan untrusted text before passing it to an LLM.

## Links
- Registry page: https://www.getdrio.com/mcp/io-github-corporatelad-promptscan
- Repository: https://github.com/corporatelad/prompt-injection-firewall

## Install
- Endpoint: https://promptscan.dev/mcp/
- Auth: Not captured

## Setup notes
- Remote endpoint: https://promptscan.dev/mcp/

## Tools
- scan_text - 
    Scan text for prompt injection attacks.

    Returns a JSON object with:
    - injection_detected (bool): true if an injection was found
    - score (float): risk score 0.0–1.0
    - label (str): "injection" or "safe"
    - attack_type (str | null): detected attack category, if any
    - layer_triggered (str | null): detection layer that flagged the text
    - latency_ms (int): detection latency in milliseconds

    Args:
        text: The text to scan (user input, retrieved doc, tool output, etc.)
        sensitivity: "low" | "medium" | "high" (default "medium")
        api_key: Your PromptScan API key (pif_...). Optional — falls back to
                 the X-API-Key header passed by the MCP client.
     Endpoint: https://promptscan.dev/mcp/

## Resources
Not captured

## Prompts
Not captured

## Metadata
- Owner: io.github.corporatelad
- Version: 0.1.0
- Runtime: Streamable Http
- Transports: HTTP
- License: Not captured
- Language: Not captured
- Stars: Not captured
- Updated: Apr 14, 2026
- Source: https://registry.modelcontextprotocol.io