CVE-2026-7147: JoeCastrom mcp-chat-studio SSRF Vulnerability Publicly Exploitable

A critical server-side request forgery (SSRF) vulnerability, identified as CVE-2026-7147, has been discovered in JoeCastrom mcp-chat-studio up to version 1.5.0. According to the National Vulnerability Database, this flaw resides within the LLM Models API component, specifically affecting the server/routes/llm.js file. Manipulation of the req.query.base_url argument allows for remote exploitation, enabling attackers to force the server to make requests to arbitrary domains.

The National Vulnerability Database highlights that the exploit for this vulnerability is now publicly available, significantly increasing the risk of widespread attacks. Despite early notification through an issue report, the project maintainers have not yet responded or provided a patch. With a CVSS score of 7.3 (HIGH), this issue represents a serious threat, potentially leading to information disclosure, port scanning of internal networks, or even access to internal services.

Attackers will leverage the public exploit to pivot into internal infrastructure, bypassing perimeter defenses. Defenders should assume active exploitation is imminent, especially given the lack of a vendor response. The attacker’s calculus here is simple: unpatched, exposed systems are low-hanging fruit for initial access and reconnaissance.