Flowise is a drag-and-drop interface for building customized large language model flows. In versions before 3.1.3, the CSV Agent node's run method improperly controls code generation by using untrusted input to construct an LLM prompt. The resulting Python code is validated by validatePythonCodeForDataFrame but this validation can be bypassed, allowing execution of arbitrary Python scripts in an unsandboxed Pyodide environment. This leads to code injection (CWE-94) and arbitrary code execution in the context of the service account. The vulnerability is addressed in Flowise version 3.1.3.

An attacker can exploit this vulnerability to execute arbitrary Python code within the Flowise environment without sandbox restrictions, potentially compromising the service account and any resources accessible to it. This could lead to unauthorized actions, data exposure, or further system compromise. The CVSS 4.0 base score is 9.5, indicating a critical severity with network attack vector, high complexity, and no privileges or user interaction required.

Upgrade Flowise to version 3.1.3 or later, where this vulnerability is fixed. No other mitigation is indicated or required as the fix addresses the root cause of the code injection vulnerability.