“Did I give you permission to delete my D:\ drive?”
Hmm… the answer here is probably YES. I doubt whatever agent he used defaulted to the ability to run all commands unsupervised.
He either approved a command that looked harmless but nuked D:\ OR he whitelisted the agent to run rmdir one day, and that whitelist remained until now.
There’s a good reason why people that choose to run agents with the ability to run commands at least try to sandbox it to limit the blast radius.
This guy let an LLM raw dog his CMD.EXE and now he’s sad that it made a mistake (as LLMs will do).
Next time, don’t point the gun at your foot and complain when it gets blown off.
The user explained what exactly went wrong later on. The AI gave a list of instructions as steps, and one of the steps was deleting a specific Node.js folder on that D:\ drive. The user didn’t want to follow the steps and just said “do everything for me” which the AI prompted for confirmation and received. The AI then indeed ran commands freely, with the same privilege as the user, however this being an AI the commands were broken and simply deleted the root of the drive rather than just one folder.
So yes, technically the AI didn’t simply delete the drive - it asked for confirmation first. But also yes, the AI did make a dumb mistake.
“Did I give you permission to delete my D:\ drive?”
Hmm… the answer here is probably YES. I doubt whatever agent he used defaulted to the ability to run all commands unsupervised.
He either approved a command that looked harmless but nuked D:\ OR he whitelisted the agent to run rmdir one day, and that whitelist remained until now.
There’s a good reason why people that choose to run agents with the ability to run commands at least try to sandbox it to limit the blast radius.
This guy let an LLM raw dog his CMD.EXE and now he’s sad that it made a mistake (as LLMs will do).
Next time, don’t point the gun at your foot and complain when it gets blown off.
The user explained what exactly went wrong later on. The AI gave a list of instructions as steps, and one of the steps was deleting a specific Node.js folder on that D:\ drive. The user didn’t want to follow the steps and just said “do everything for me” which the AI prompted for confirmation and received. The AI then indeed ran commands freely, with the same privilege as the user, however this being an AI the commands were broken and simply deleted the root of the drive rather than just one folder.
So yes, technically the AI didn’t simply delete the drive - it asked for confirmation first. But also yes, the AI did make a dumb mistake.