Code Injection in PraisonAI deploy/api.py agents_file Handling

Successful exploitation allows arbitrary Python code execution in the context of the PraisonAI deployment workflow. This can lead to full compromise of the affected application environment, including unauthorized access to sensitive data, modification of application behavior or generated code, installation of persistence mechanisms, lateral movement from the compromised host, and disruption or destruction of services. The available scoring indicates high impact to confidentiality, integrity, and availability, with scope change.

If immediate upgrade is not possible, restrict access to the vulnerable deployment functionality to only fully trusted administrators and isolate the service from untrusted users and networks. Disable or avoid use of the affected code path that processes the agents_file parameter until patched. Apply strict allowlist validation for accepted parameter values, reject characters or constructs that could alter Python syntax, and run the deployment component in a tightly sandboxed environment with minimal operating system and application privileges. Monitor for anomalous subprocess creation and unexpected code generation activity associated with the deployment workflow.

Upgrade PraisonAI to version 4.6.78 or later, which is listed as unaffected. Review and refactor any code paths that generate executable Python code from user-controllable input, particularly the handling of the agents_file parameter in deploy/api.py. Replace direct string interpolation into executable code with safe parameter handling, strict input validation, and designs that avoid dynamic code generation where possible. After upgrading, inspect deployment workflows and generated artifacts for signs of prior abuse and rotate any credentials that may have been exposed to the vulnerable execution context.