CVE-2026-80347: Server-Side Request Forgery (SSRF) in kazuph mcp-fetch

The mcp-fetch tool performs SSRF protection by checking fetch targets against a guard that validates IP addresses. However, it fails to remove brackets surrounding IPv6 literals before validation. When an IPv6 literal such as http://[::1]/ is checked, the hostname retains brackets, causing the net.isIP function to return zero, skipping private address checks. The guard then attempts DNS resolution, which fails, marking the target as safe. The HTTP client later removes the brackets and connects to the actual IP address, allowing access to loopback and private IPv4-mapped addresses that should have been blocked. This flaw enables attackers controlling fetch targets to read internal network responses via SSRF.

An attacker able to influence the fetch target argument can bypass SSRF protections and access internal network resources, including loopback and private IP addresses. This can lead to unauthorized disclosure of sensitive internal data by reading internal responses into the model context. The vulnerability has a high severity score (CVSS 8.7) indicating significant risk if exploited.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until a fix is available, users should avoid exposing the fetch target argument to untrusted input or implement additional network-level restrictions to prevent unauthorized internal requests.