AIKIDO-2026-377750
strands-agents-tools is vulnerable to Improper Neutralization of Input Used for LLM Prompting
81
High Risk
This Affects:
strands-agents-toolsTL;DR
The python_repl tool requires operator approval through a human consent gate before executing Python on the agent host. When both the batch and python_repl tools are registered on the same agent, a crafted prompt can pass non_interactive_mode as a keyword argument through the batch tool, causing python_repl to run without triggering consent. This enables arbitrary Python code execution on the host. The fix makes non_interactive_mode read only from the STRANDS_NON_INTERACTIVE environment variable so it cannot be set through prompt input.
Who does this affect?
You are affected if you are using a version that falls within the vulnerable range and you register both the batch and python_repl tools on an agent that processes untrusted content.
Background info
strands-agents-tools is vulnerable to Improper Neutralization of Input Used for LLM Prompting in versions 0.0.1 - 0.8.4.
How to fix this
Upgrade the strands-agents-tools library to the patch version.
Links
GitHub Release
Are You Affected?
Connect your repositories to instantly see whether vulnerable or malicious packages exist in your codebase.
Free. No credit card required.
I consent to receiving marketing communications based on Aikido’s Privacy Policy.
SOC 2Compliant
ISO 27001CompliantISO 42001Compliant
FedRAMPImplementing© All Intel data is openly available and commercially licensed.