Hi guys, I’m having this problem with the Execute command tool connected to an AI agent, basically when the tool returns a very large output, the AI Agent crashes and returns this error message:
Problem in node ‘ReconAI‘
Cannot read properties of undefined (reading ‘content’)
can anyone help me figure out how to solve it? if i have to link it to another tool to split the output into chunks, can you tell me the name? unfortunately, i couldn’t find anything related to this one.
Hey @Saverio_Valentino ,
What is the command you are trying to run in this node ?
Hi @mrrobot , it is not so much the command with which I have the problem, but it is the response, an example is like when running the command whois facebook.com.
I would also like to point out for those who will see this question and are not familiar with cybersecurity (this is definitely not aimed at you @mrrobot 
), that the purposes are purely academic, I am working on a university project in cybersecurity and the tests are being carried out on metasploitable 2, the command whois facebook.com is an example to trigger the error, moreover it is a legitimate command, because the information is public.
Hey @Saverio_Valentino ,
If you’re using Docker you have to build the image with command you want to run
yes I am using docker and the command exists and is executed, the output of the command when returned to the AI Agent gives me that error.
Another small problem I noticed this morning, sometimes the Agent executes the commands all at once, and sometimes it executes them one at a time, is there any way to always have it execute the commands one at a time ? does this just need to be specified in the prompt or is there any option ?