Nezha, a network monitoring tool, contains a host header injection flaw in its mintTransferTool component. The flaw allows the service to build fs.download_url and fs.upload_url transfer URLs using the unvalidated Host header from a victim’s MCP request. By crafting a Host header, an attacker can redirect file uploads to a malicious server or capture one-time download tokens and replay them, thereby exfiltrating files from the system.

Systems running Nezha versions 2.1.0 through 2.3.14 are vulnerable. The affected product is nezhahq:nezha. The vulnerability is present in any installation of Nezha before the 2.3.15 release, irrespective of additional configurations.

The CVSS score of 6.3 classifies the flaw as medium severity. The EPSS score is < 1%, indicating a very low probability of exploitation, and the vulnerability is not listed in the CISA Known Exploited Vulnerabilities catalog, so no publicly disclosed exploits are known. However, the attack requires control over the HTTP Host header seen by the MCP service, which can be achieved by an attacker who can influence network traffic to the host or compromise the MCP client configuration. Once the Host header is injected, the attacker can route upload traffic elsewhere or replay download tokens to exfiltrate arbitrary files.