Public Docs / MCP / Overview
MCP Overview
Phoenix Security exposes its vulnerability intelligence platform over the Model Context Protocol (MCP), letting AI assistants query CVE data, risk scores, EOL status, and supply chain signals in real time.
What is MCP?
The Model Context Protocol is an open standard that connects AI models to external data sources and tools using a JSON-RPC 2.0 message format. When you configure an MCP server in your AI client, the model can call tools and read structured resources during a conversation — no copy-paste, no manual API calls.
Phoenix implements MCP as an HTTP endpoint that accepts JSON-RPC messages. A lightweight stdio bridge (included in your account resources) translates between the client's stdin/stdout and the Phoenix HTTP endpoint, so the integration looks native in every supported client.
Endpoint
- Production:
https://phxintel.security/api/v1/mcp - Claude Desktop alias:
https://phxintel.security/api/v1/mcp/claude - ChatGPT Desktop alias:
https://phxintel.security/api/v1/mcp/chatgpt - Protocol: JSON-RPC 2.0
- Method: HTTP POST with
Content-Type: application/json
Authentication
All MCP requests require an API key passed in the x-api-key HTTP header.
You generate API keys from your Phoenix account settings page.
Keys used for MCP must carry one of the following scopes:
mcp— dedicated MCP scope (Registered tier)api_power— power-user scope (Pro tier)api_integration— integration scope (Pro tier)api_unlimited— unlimited scope (Enterprise tier)
Key prefixes: Registered keys begin with phx_reg_, Pro keys with phx_pro_,
and Enterprise keys with phx_ent_. The scope is enforced server-side — passing a key
from the wrong tier returns an HTTP 403 or a JSON-RPC error.
Access Tiers
The MCP server applies the same tier-based access model as the REST API. Your API key's scope determines which tools are visible and how much detail is returned in each response.
| Tier | Key Scope | Tools Available | Score Detail |
|---|---|---|---|
| Registered | mcp |
Standard tools | Score, tier, top 3 generic reasons |
| Pro | api_power / api_integration |
Standard tools | Score + component levels (H/M/L), top 5 specific reasons |
| Enterprise | api_unlimited |
All tools including simulation | Full numeric component values, rationale, executive summary |
What Can You Do?
- Ask your AI assistant to look up any CVE and get Phoenix scoring, EPSS, KEV status, and threat actor associations.
- Check whether a vendor/product combination is enterprise-critical and get its risk classification.
- Query the CISA Known Exploited Vulnerabilities catalog.
- Get end-of-life risk scores and replacement recommendations for products in your environment.
- Search for high-profile CVEs by PS-HP tier (confirmed, likely, emerging).
- Evaluate packages against supply chain firewall rules (Enterprise, when feature is enabled).
Next Steps
- Installation Guide — configure Phoenix MCP in Claude Desktop, ChatGPT Desktop, or a custom client
- Tools Reference — full tool catalog with input schemas and tier requirements
- Resources Reference — read-only resource URIs