An unauthenticated vulnerability in the MCP Documentation Server allows unauthorized network users to manipulate the AI knowledge base and access sensitive documents.

The application incorrectly binds the web interface to all network interfaces instead of localhost, and fails to implement authentication for the document management API. This allows any attacker on the local network, VPN, or connected bridge to perform unauthorized CRUD operations on the corpus.

This vulnerability carries a CVSS score of 8.8, reflecting its high severity. Successful exploitation could lead to the total compromise of the AI assistant's knowledge base, leading to data exfiltration, the injection of malicious information, or the destruction of critical documentation. This poses a significant risk to organizational data integrity and confidentiality.

Immediate Action: Update the mcp-documentation-server package to version 1.13.1 or later immediately.

Proactive Monitoring: Monitor network traffic for unauthorized access attempts directed at port 3080 and review access logs for unexpected API calls to the /api/documents or /api/search-all endpoints.

Compensating Controls: Restrict network access to the server by using host-based firewalls to ensure the service is only reachable from authorized management workstations.

Public Exploit Available: Yes, a published PoC exists.

Given the ease of access and the potential for severe data tampering, organizations should prioritize patching this vulnerability immediately. Ensure that the updated version is deployed across all environments and verify that the service is no longer listening on public or unauthorized network interfaces.