AIKIDO-2026-985837

praisonaiagents is vulnerable to Code Injection

Code InjectionCVE-2026-61447 Published Aug 13, 2026

100

Critical Risk

This Affects:

praisonaiagents

0.13.21 - 1.6.77

Fixed in 1.6.78

Are you affected? Scan for Free

TL;DR

CodeAgent._execute_python() writes LLM-generated Python to a temp file and runs it in a subprocess with os.environ.copy(), with no AST validation, import restrictions, or sandbox enforcement—even when CodeConfig(sandbox=True) is set. An attacker who can influence the generated code through prompt injection can read environment secrets such as API keys and cloud credentials and run arbitrary code on the host. The fix enforces sandboxing for code execution and stops passing the full parent environment into the subprocess.

Who does this affect?

You are affected if you are using a version that falls within the vulnerable range and use CodeAgent (or equivalent code-execution tooling) with LLM-generated or otherwise untrusted code.