OpenCode Flaw Lets Malicious MCP Servers Execute Code During Sign-In on Windows and WSL
Summary
A vulnerability in the OpenCode AI coding agent has been patched. The flaw allowed malicious or hijacked MCP servers to execute arbitrary code on a developer's Windows or WSL machine during sign-in, with the same privileges as the developer.
IFF Assessment
This vulnerability allows attackers to execute code on developer machines, posing a direct threat to their systems and potentially leading to further compromise.
Severity
The vulnerability allows for remote code execution on the developer's machine, with a high impact on confidentiality, integrity, and availability, and requires low complexity to exploit.
Defender Context
Developers using AI coding agents like OpenCode should ensure they are running the latest versions to mitigate risks. This incident highlights the need for security scrutiny of third-party integrations within development tools, as malicious servers could leverage these connections for attacks.