OpenCode Flaw Lets Malicious MCP Servers Execute Code During Sign-In on Windows and WSL

A vulnerability in the OpenCode AI coding agent enabled malicious MCP servers to execute code on a developer's Windows or WSL machine during the sign-in process. This code would run with the developer's own access level due to OpenCode's connection to MCP servers for integrating external tools. The company behind OpenCode, Anomaly, addressed the issue in version 1.18.33. The fix was implemented to prevent malicious or hijacked MCP servers from exploiting the flaw. The vulnerability posed a risk to developers using OpenCode on Windows and WSL machines. Anomaly released the patch to mitigate this security risk.