Pangea MCP Server – AI Security Services Access
Pangea MCP Server is an official Pangea Cyber Model Context Protocol server that exposes selected Pangea security services to MCP-compatible AI clients. It lets agents use Pangea APIs for AI Guard, threat intelligence, Redact, Secure Audit Log, Vault, and related security workflows while relying on Pangea tokens and service-level permissions.
Overview
Pangea MCP Server connects AI agents to Pangea Cyber's security APIs through the Model Context Protocol. It is designed for security-oriented agent workflows where a model needs live access to trusted services for prompt protection, threat intelligence lookups, sensitive-data redaction, audit logging, and vault-backed secret retrieval. The server is distributed as the @pangeacyber/mcp-server package and can run as a local stdio server or as a streamable HTTP server.
What the MCP server enables
The server exposes tools for several Pangea services. AI Guard can analyze and redact text to reduce prompt manipulation and unsafe data transfer. Domain Intel, IP Intel, URL Intel, and File Intel provide reputation, WHOIS, geolocation, proxy, VPN, embargo, and hash reputation workflows. Redact can remove sensitive information from supplied text. Secure Audit Log tools can create and search tamper-evident audit entries. Vault tools can list, retrieve, delete, and generate vault items such as secrets, keys, tokens, and folders.
When to use it
Use this MCP server when an AI agent needs security context that should come from live Pangea APIs instead of static model knowledge. Examples include checking whether an IP address is associated with a VPN or sanctioned geography, redacting personal information from text before it is passed onward, recording agent actions in Secure Audit Log, or retrieving controlled secret metadata from Vault. It is most useful for security operations, compliance automation, AI safety guardrails, and developer workflows that need Pangea services inside an MCP client.
Connection and authentication
For local clients such as Claude Desktop, the documented connection is stdio using npx -y @pangeacyber/mcp-server. The required environment variables are PANGEA_VAULT_TOKEN, PANGEA_VAULT_ITEM_ID, and PANGEA_AUDIT_CONFIG_ID. The Vault token is used to fetch a stored API token for the other Pangea services at runtime. The server can also run over streamable HTTP with --transport httpStream --port 8080, and Pangea AuthN can optionally protect that HTTP server with issuer, client ID, and client secret environment variables.
Key considerations
The GitHub repository is archived and read-only, so teams should review release status and maintenance expectations before production use. The server requires Node.js 24.11.0 or newer. Tokens must be scoped only to the Pangea services that the agent needs, and Vault-backed token handling should be treated as sensitive infrastructure. Streamable HTTP deployments should not be exposed publicly without authentication such as the documented Pangea AuthN option.
Supported Transports
stdio
Command: npx
Args:
-y@pangeacyber/mcp-server
streamable_http
URL: http://localhost:8080/mcp
Frequently Asked Questions
- When should an AI agent use the Pangea MCP Server?
- Use it when an agent needs live Pangea security services, such as checking domain, IP, URL, or file reputation, redacting sensitive text, applying AI Guard checks, writing audit events, or working with Vault-backed items.
- What does this MCP server add to an AI agent's capabilities?
- It adds structured access to Pangea security APIs, including AI Guard, Domain Intel, Embargo, File Intel, IP Intel, Redact, Secure Audit Log, URL Intel, and Vault tools, so the agent can use current service results instead of relying only on model knowledge.
- What can the AI agent access or manage through this server?
- Depending on the configured Pangea token permissions, the agent can run reputation and WHOIS lookups, check embargo data, detect proxy or VPN IPs, geolocate IP addresses, redact sensitive text, create or search audit log entries, and retrieve or manage selected Vault items.
- How is authentication configured for this MCP server?
- The stdio server uses `PANGEA_VAULT_TOKEN`, `PANGEA_VAULT_ITEM_ID`, and `PANGEA_AUDIT_CONFIG_ID`. The Vault token fetches a stored token that has access to the enabled Pangea services. Streamable HTTP deployments can optionally add Pangea AuthN using issuer, client ID, and client secret environment variables.
- Which transport should be used for this MCP server?
- Use `stdio` for local desktop or IDE MCP clients with the documented `npx` command. Use `streamable_http` when running the server as an HTTP service, typically on a controlled port such as 8080 and preferably protected with the documented Pangea AuthN option.