Browse / MCP and Tool Server Flaws / CVE-2025-58444
MCP Inspector is Vulnerable to Potential Command Execution via XSS When Connecting to an Untrusted MCP Server
- Class
- MCP and Tool Server Flaws (filed by a model that read the advisory)
- Technique
- Code and command execution
- Affected
- @modelcontextprotocol/inspector (npm)
- Gets past
- Input filter or sanitiser
- CVSS
- CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
- EPSS
- 0.7% chance of exploitation in 30 days
How it works, in short
The inspector renders a redirect URL from an untrusted MCP server without sanitising it, allowing script injection that reaches the local proxy. (Our summary, written by a model. The source text is below.)
Details
An XSS flaw exists in the MCP Inspector local development tool when it renders a redirect URL returned by a remote MCP server. If the Inspector connects to an untrusted server, a crafted redirect can inject script into the Inspector context and, via the built-in proxy, be leveraged to trigger arbitrary command execution on the developer machine. Version 0.16.6 hardens URL handling/validation and prevents script execution.
> Thank you to the following researchers for their reports and contributions:
> * Raymond (Veria Labs)
> * Gavin Zhong, <superboyzjc@gmail.com> & Shuyang Wang, <swang@obsidiansecurity.com>.
References
- https://github.com/modelcontextprotocol/inspector/security/advisories/GHSA-g9hg-qhmf-q45m
- https://nvd.nist.gov/vuln/detail/CVE-2025-58444
- https://github.com/modelcontextprotocol/inspector/commit/650f3090d26344a672026b737d81586595bb1f60
- https://github.com/modelcontextprotocol/inspector
- https://www.npmjs.com/package/@modelcontextprotocol/inspector/v/0.16.6
Source: GitHub Advisory Database, used under CC BY 4.0.