PraisonAI's MCP command handling function parse_mcp_command() suffers from an OS command injection vulnerability (CVE-2026-41497). The function lacks proper validation of executables and arguments, allowing arbitrary commands such as bash, python, or /bin/sh with inline code execution to be executed via subprocess calls. Although a patch (commit 47bff654) was released, it does not implement critical controls like an allowlist of permitted commands, basename checks, or argument inspection, leaving the vulnerability partially unmitigated. Versions prior to 4.5.149 are affected, with 4.5.149 containing the fix. The vulnerability has a CVSS score of 9.8, indicating critical severity.