pan-mcp-relay

pan-mcp-relay

Security-enhanced MCP Relay/Proxy server providing real-time AI threat protection for MCP Clients, scanning tool descriptions, call parameters, and responses for threats like prompt injections and malicious URLs.

Category
Visit Server

README

Prisma AIRS MCP Security Relay

By Palo Alto Networks

<!--TOC-->

<!--TOC-->

<a id="overview" href="#overview">

Overview

</a>

pan-mcp-relay is a security-enhanced Model Context Protocol (MCP) Relay (Proxy) Server providing real-time AI threat protection for MCP Clients, built with the Prisma AIRS AI Runtime API Intercept.

pan-mcp-relay will help protect MCP Clients such as IDE's, LLM Chat Clients, and AI Agents from harmful MCP Server Tools by automatically scanning and blocking various threats, including prompt injections, malicious URLs, insecure outputs, AI agentic threats, sensitive data loss, and more.

The MCP Relay scans all MCP Server tool descriptions, tool call parameters, and tool call responses.

For licensing, onboarding, activation, and to obtain an API authentication key and profile name, refer to the Prisma AIRS AI Runtime: API Intercept Overview administration documentation.

Prerequisites

Follow the following Activation and Onboarding guides:

  1. Create and associate a deployment profile for Prisma AIRS AI Runtime API intercept in your Customer Support Portal.
  2. Onboard Prisma AIRS AI Runtime API intercept in Strata Cloud Manager.
  3. Manage applications, API keys, security profiles, and custom topics in Strata Cloud Manager.

Requirements for Prisma AIRS API Usage

  1. API Key: Generate an API Key during the onboarding process in Strata Cloud Manager (see the onboarding prerequisite step above).
  • Specify the API Key via configuration flag, environment variable or configuration file.
  1. AI Security Profile Name: This is the API security profile you created during the onboarding process in Strata Cloud Manager (see the prerequisite step on creating an API security profile above).
  • Specify the profile name or the profile ID via configuration flag, environment variable, or configuration file.

[!NOTE] You can manage API keys and AI security profiles in Strata Cloud Manager.

  1. Log in to Strata Cloud Manager.
  2. Navigate to Insights > Prisma AIRS > Prisma AIRS AI Runtime: API Intercept.
  3. In the top right corner, click:
  • Manage > API Keys to copy, regenerate, or rotate the API key token.
  • Manage > Security Profiles to fetch details or update AI security profiles.
  • Manage > Custom Topics create or update custom topics for custom topic guardrails threat detections.

For complete details, refer to the Activation and Onboarding guides on how to Manage Applications, API Keys, Security Profiles, and Custom Topics.

<a id="installation" href="#installation">

Installation

</a>

We highly recommend using uv over pip or pipx. Try it! You'll love it.

<details><summary>Installing uv</summary>

See more installation methods at https://docs.astral.sh/uv/getting-started/installation/.

Standalone installer

uv provides a standalone installer to download and install uv:

=== "macOS and Linux"

Use `curl` to download the script and execute it with `sh`:

```console
$ curl -LsSf https://astral.sh/uv/install.sh | sh
```

If your system doesn't have `curl`, you can use `wget`:

```console
$ wget -qO- https://astral.sh/uv/install.sh | sh
```

Request a specific version by including it in the URL:

```console
$ curl -LsSf https://astral.sh/uv/0.8.13/install.sh | sh
```

=== "Windows"

Use `irm` to download the script and execute it with `iex`:

```pwsh-session
PS> powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
```

Changing the [execution policy](https://learn.microsoft.com/en-us/powershell/module/microsoft.powershell.core/about/about_execution_policies?view=powershell-7.4#powershell-execution-policies) allows running a script from the internet.

Request a specific version by including it in the URL:

```pwsh-session
PS> powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/0.8.13/install.ps1 | iex"
```

!!! tip

The installation script may be inspected before use:

=== "macOS and Linux"

    ```console
    $ curl -LsSf https://astral.sh/uv/install.sh | less
    ```

=== "Windows"

    ```pwsh-session
    PS> powershell -c "irm https://astral.sh/uv/install.ps1 | more"
    ```

Alternatively, the installer or binaries can be downloaded directly from [GitHub](#github-releases).

</details>

Update uv if already installed - uv releases updates frequently, so make sure you're on the latest.

uv self update

<a id="quickstart-install-and-run-via-pypi-org" href="#quickstart-install-and-run-via-pypi-org">

Quickstart: Install and Run via pypi.org

</a>

Install and run pan-mcp-relay in one line, showing the CLI help:

uvx pan-mcp-relay@latest --help

[!NOTE] While this makes it easy to get started, you can get faster startup times by following the Standard Install

<a id="standard-install" href="#standard-install">

Standard Install

</a>

uv installs to a standard location, which automatically added to your $PATH during uv installation.

On macOS and Linux, this is usually ~/.local/bin.

# Installs to $(uv tool dir)/pan-mcp-relay
uv tool install pan-mcp-relay@latest

# Executable symlinked to $(uv tool dir --bin)/pan-mcp-relay
pan-mcp-relay --help

<a id="install-the-bleeding-edge-development-version" href="#install-the-bleeding-edge-development-version">

Install the bleeding edge development version

</a>

uv tool install https://github.com/PaloAltoNetworks/aisecurity-mcp-relay.git
pan-mcp-relay --help

<a id="upgrade" href="#upgrade">

Upgrade

</a>

If you're using uvx pan-mcp-relay@latest, you're already using the latest version.

Update the installed version with

uv tool install --upgrade pan-mcp-relay
  1. Install or Upgrade pan-mcp-relay using uv: uv tool install --upgrade pan-mcp-relay
  2. Create an mcp-relay.yaml configuration file containing:
    1. Prisma AIRS API Configuration
      • Prisma AIRS API Key
      • Prisma AIRS AI Profile
      • Prisma AIRS API Endpoint (Optional, defaults to United States Prisma AIRS API Endpoint.)
    2. MCP Servers
      • Supports stdio, SSE, and Streamable HTTP MCP Servers.
    3. Run the MCP Relay Server

<a id="configuration" href="#configuration">

Configuration

</a>

<a id="relay-configuration" href="#relay-configuration">

MCP Relay Server Configuration

</a>

<a id="configuration-precedence" href="#configuration-precedence">

Configuration Precedence

</a>

  1. CLI Flags
  2. Environment Variables
  3. .env file variables
  4. Configuration File

pan-mcp-relay supports loading a configuration file via --config-file flag, MCP_RELAY_CONFIG_FILE environment variable, or detecting a configuration file on a set of pre-determined locations:

  • --config-file (-c) CLI flag
  • MCP_RELAY_CONFIG_FILE environment variable
  • ./mcp-relay.yaml
  • ~/.mcp-relay.yaml
  • ~/.config/pan-mcp-relay/mcp-relay.yaml
  • ./mcp-relay.json
  • ~/.mcp-relay.json
  • ~/.config/pan-mcp-relay/mcp-relay.json

<a id="configuration-file-format" href="#configuration-file-format">

Configuration File Format

</a>

An example MCP Relay Server configuration is available in the repository in examples/config/mcp-relay.yaml

Copy or create a new mcp-relay.yaml file. The only required section is mcpServers: {}.

# mcpRelay section is optional.
# Specify API Key, AI Profile and API Endpoint via CLI flags or environment variables.
mcpRelay:
  # Prisma AIRS API Key (required), environment variables supported
  apiKey: |
    ${PRISMA_AIRS_API_KEY}
  # Prisma AIRS AI Profile Name or ID (required), environment variables supported
  aiProfile: |
    your-ai-profile-name-or-id
  # Endpoint Optional, default shown
  # endpoint: |
  #  https://service.api.aisecurity.paloaltonetworks.com

# mcpServers section requires at least one MCP Server.
mcpServers:
  # Example / demo MCP Servers
  homebrew:
    command: brew
    args:
      - mcp-server
    # Optional working directory, supports environment variables
    cwd: "${HOMEBREW_PREFIX}"
    # optional environment variables, inherited from the parent shell
    env:
      HOMEBREW_NO_ENV_HINTS: 1
  fetch:
    command: uvx
    args:
      - mcp-server-fetch

  # filesystem:
  #   command: npx
  #   args:
  #     - -y
  #     - '@modelcontextprotocol/server-filesystem'
  #     - /var/tmp/change/this/path
  #     # Command line arguments support environment variables
  #     - "${XDG_PICTURES_DIR}"

<a id="environment-variables" href="#environment-variables">

Environment Variables

</a>

In addition to CLI flags and the configuration file, pan-mcp-relay supports setting configuration through the following environment variables:

# Required for Prisma AIRS API
PRISMA_AIRS_API_KEY=YOUR_API_KEY
# Required for Prisma AIRS API
PRISMA_AIRS_AI_PROFILE=YOUR_AI_PROFILE_NAME

# Optional, default is https://service.api.aisecurity.paloaltonetworks.com
PRISMA_AIRS_API_ENDPOINT=https://service.api.aisecurity.paloaltonetworks.com
# See https://pan.dev/prisma-airs/scan/api/#scan-api-endpoints for additional regional API endpoints

# Defaults shown for all other Environment Variables
MCP_RELAY_CONFIG_FILE=~/.config/pan-mcp-relay/mcp-relay.yaml
MCP_RELAY_TRANSPORT=stdio
# Host For SSE Transport Mode
MCP_RELAY_HOST=127.0.0.1
# Port for SSE Transport Mode
MCP_RELAY_TOOL_CACHE_TTL=86400
MCP_RELAY_MAX_SERVERS=32
MCP_RELAY_MAX_TOOLS=256
# Path to optional .env file
MCP_RELAY_DOTENV=
# Supports $PATH-style colon-separated list and environment variables. Directory entries will search for a file named `.env`
# MCP_RELAY_DOTENV=$HOME/.env:~/.config/pan-mcp-relay:$PWD

MCP_RELAY_LOG_LEVEL=INFO

<a id="mcp-client-configuration" href="#mcp-client-configuration">

MCP Client Configuration

</a>

Create a new config file mcp-relay.json to use with your MCP Client (IDE, Chat Client, or Agent).

{
  "mcpServers": {
    "pan-mcp-relay": {
      "command": "uvx",
      "args": [
        "pan-mcp-relay@latest"
      ]
    }
  }
}

For faster startup times, install/upgrade pan-mcp-relay using uv tool install --upgrade pan-mcp-relay, and use the command name directly:

{
  "mcpServers": {
    "pan-mcp-relay": {
      "command": "pan-mcp-relay"
    }
  }
}

[!CAUTION]

pan-mcp-relay should be the only MCP Server listed in your MCP Client configuration file.

Any other MCP Servers listed here will NOT be visible to the security capabilities of the MCP Relay.

Optionally, specify additional environment variables or CLI flags:

{
  "mcpServers": {
    "pan-mcp-relay": {
      "command": "uvx",
      "args": [
        "pan-mcp-relay",
        "--config-file",
        "~/.config/pan-mcp-relay/mcp-relay.yaml"
      ],
      "env": {
        "PRISMA_AIRS_AI_PROFILE": "your-ai-profile",
        "MCP_RELAY_LOG_LEVEL": "WARNING"
      }
    }
  }
}

<a id="transport-options" href="#transport-options">

Transport Options

</a>

The MCP Relay Server supports the standard MCP transport mechanisms:

  • stdio Transport: For local process communication (default)
  • Streamable Transport: For HTTP-based communication, supporting SSE (Server-Sent Events).
  • SSE Transport: For backwards compatibility with SSE-only MCP Clients.

<a id="usage" href="#usage">

Usage

</a>

<a id="pan-mcp-relay-cli-usage" href="#pan-mcp-relay-cli-usage">

pan-mcp-relay CLI Usage

</a>

pan-mcp-relay --help
Usage: pan-mcp-relay [OPTIONS] COMMAND [ARGS]...

  Run the MCP Relay Server.

Options:
  -k, --api-key TEXT              Prisma AIRS API Key [PRISMA_AIRS_API_KEY=]
  -e, --api-endpoint TEXT         Prisma AIRS API Endpoint [PRISMA_AIRS_API_ENDPOINT=]
  -p, --ai-profile TEXT           Prisma AIRS AI Profile Name or ID [PRISMA_AIRS_AI_PROFILE=]
  -c, --config-file FILE          Path to configuration file (yaml, json) [MCP_RELAY_CONFIG_FILE=]
  -t, --transport [stdio|sse|http]
                                  Transport protocol to use [MCP_RELAY_TRANSPORT=]  [default: stdio]
  -h, --host TEXT                 Host for HTTP/SSE server [MCP_RELAY_HOST=]  [default: 127.0.0.1]
  -p, --port INTEGER              Port for HTTP/SSE server [MCP_RELAY_PORT=]  [default: 8000]
  -TTL, --tool-registry-cache-ttl INTEGER
                                  Tool registry cache TTL (in seconds) [MCP_RELAY_TOOL_CACHE_TTL=]  [default: 86400]
  -MS, --max-mcp-servers INTEGER  Maximum number of downstream MCP servers to allow [MCP_RELAY_MAX_SERVERS=]  [default: 32]
  -MT, --max-mcp-tools INTEGER    Maximum number of MCP tools to allow [MCP_RELAY_MAX_TOOLS=]  [default: 256]
  --env TEXT                      Use ./.env file, or specify colon separated path to .env file(s)or directories containing .env files. [MCP_RELAY_DOTENV=]
  --system-ca                     Use System CA instead of Mozilla CA Bundle
  --capath FILE                   Path to Custom Trusted CA bundle
  -d, --debug                     Extremely verbose logging output (DEBUG)
  -v, --verbose                   Verbose logging output (INFO, DEFAULT)
  -q, --quiet                     Minimal logging output (WARNING)
  --show-config, --dump-config    Show configuration and exit [MCP_RELAY_SHOW_CONFIG=]
  -V, --version                   Show the version and exit.
  --help                          Show this message and exit.

Commands:
  json-schema  Print Configuration File JSON Schema, optionally to a file.
  run          Run the MCP Relay Server.
  show-config  Print MCP Relay Configuration and exit.

<a id="running-the-relay-server" href="#running-the-relay-server">

Running the Relay Server

</a>

<a id="stdio-transport-default" href="#stdio-transport-default">

stdio transport (default)

</a>

# Run with stdio transport with automatic configuration detection (default)
uvx pan-mcp-relay

# Specify a configuration file
uvx pan-mcp-relay --config-file ~/.config/pan-mcp-relay/mcp-relay.yaml

# Specify a path to a custom .env file
uvx pan-mcp-relay --env /var/run/secrets/.env

<a id="streamable-http-transport" href="#streamable-http-transport">

Streamable HTTP transport

</a>

# Run with Streamable transport
pan-mcp-relay --transport=http

<a id="sse-transport" href="#sse-transport">

SSE transport

</a>

# Run with SSE transport
pan-mcp-relay --transport=sse

<a id="legal" href="#legal">

Legal

</a>

Copyright © 2025, Palo Alto Networks

Licensed under the Polyform Internal Use License 1.0.0 (the "License"); you may not use this file except in compliance with the License.

You may obtain a copy of the License at:

https://polyformproject.org/licenses/internal-use/1.0.0

(or)

https://github.com/polyformproject/polyform-licenses/blob/76a278c4/PolyForm-Internal-Use-1.0.0.md

As far as the law allows, the software comes as is, without any warranty or condition, and the licensor will not be liable to you for any damages arising out of these terms or the use or nature of the software, under any kind of legal claim.

<!---Protected_by_PANW_Code_Armor_2024 - Y3ByfC9haWZ3L29wZW5zb3VyY2UvYWlzZWN1cml0eS1tY3AtcmVsYXl8MjE5NDV8bWFpbg== --->

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured