Microsoft Copilot Cowork Exfiltrates Files

#1

Post by Kneenex » Mon, May 25, 2026, 9:45 PM UTC

Microsoft Copilot Cowork Exfiltrates Files
promptarmor.com

Re: Microsoft Copilot Cowork Exfiltrates Files

#2

Post by 2001zhaozhao » Mon, May 25, 2026, 10:36 PM UTC

AKA, if a malicious skill got into your AI agent, you're cooked.

I think this isn't surprising, nor do I think it should be considered a prompt injection at all. An AI skill is akin to a plugin for traditional software - if you install a malicious IDE extension or Outlook plugin, the attacker can also do whatever they want to the PC and exfiltrate whatever data they want to. So this article is a big nothingburger.

Re: Microsoft Copilot Cowork Exfiltrates Files

#3

Post by hansmayer » Mon, May 25, 2026, 10:39 PM UTC

Well, isn't that swell - good that meanwhile countless MBA cretins have "adopted" enterprise-wide Copilot integrations, to make their companies "AI native" or whatever the word is on LinkedinLunatics street these days.

Re: Microsoft Copilot Cowork Exfiltrates Files

#4

Post by Jabrov » Mon, May 25, 2026, 10:39 PM UTC

2001zhaozhao wrote: post #2

AKA, if a malicious skill got into your AI agent, you're cooked. I think this isn't surprising, nor do I think it should be considered a prompt injection at all. An AI skill is akin to a plugin for traditional software - if you install a malicious IDE extension or Outlook plugin, the attacker can also do whatever they want to the PC and exfiltrate whatever data they want to. So this article is a big nothingburger.

It's yet another surface for dependency attacks

Re: Microsoft Copilot Cowork Exfiltrates Files

#5

Post by aabhay » Mon, May 25, 2026, 10:40 PM UTC

2001zhaozhao wrote: post #2

AKA, if a malicious skill got into your AI agent, you're cooked. I think this isn't surprising, nor do I think it should be considered a prompt injection at all. An AI skill is akin to a plugin for traditional software - if you install a malicious IDE extension or Outlook plugin, the attacker can also do whatever they want to the PC and exfiltrate whatever data they want to. So this article is a big nothingburger.

Its actually even worse — its advertising for their product

Re: Microsoft Copilot Cowork Exfiltrates Files

#6

Post by nico » Mon, May 25, 2026, 10:40 PM UTC

2001zhaozhao wrote: post #2

AKA, if a malicious skill got into your AI agent, you're cooked. I think this isn't surprising, nor do I think it should be considered a prompt injection at all. An AI skill is akin to a plugin for traditional software - if you install a malicious IDE extension or Outlook plugin, the attacker can also do whatever they want to the PC and exfiltrate whatever data they want to. So this article is a big nothingburger.

I wonder if via-skill could become a software distribution channel. A bit like what has happened with LLM wiki

Re: Microsoft Copilot Cowork Exfiltrates Files

#7

Post by cyanydeez » Mon, May 25, 2026, 10:43 PM UTC

2001zhaozhao wrote: post #2

AKA, if a malicious skill got into your AI agent, you're cooked. I think this isn't surprising, nor do I think it should be considered a prompt injection at all. An AI skill is akin to a plugin for traditional software - if you install a malicious IDE extension or Outlook plugin, the attacker can also do whatever they want to the PC and exfiltrate whatever data they want to. So this article is a big nothingburger.

ai skill is not just a plugin. given the right model, supposedly, it can do much more. since everyones harness tends to be tied to the model, it has a whole tool set to use.

Re: Microsoft Copilot Cowork Exfiltrates Files

#8

Post by bestony » Mon, May 25, 2026, 10:43 PM UTC

[flagged]

Re: Microsoft Copilot Cowork Exfiltrates Files

#9

Post by bestony » Mon, May 25, 2026, 10:44 PM UTC

Large-scale adoption will take time; we still need a lot more infrastructure, such as security, auditing, and payment systems.

Re: Microsoft Copilot Cowork Exfiltrates Files

#10

Post by 0gs » Mon, May 25, 2026, 10:47 PM UTC

2001zhaozhao wrote: post #2

AKA, if a malicious skill got into your AI agent, you're cooked. I think this isn't surprising, nor do I think it should be considered a prompt injection at all. An AI skill is akin to a plugin for traditional software - if you install a malicious IDE extension or Outlook plugin, the attacker can also do whatever they want to the PC and exfiltrate whatever data they want to. So this article is a big nothingburger.

i think people are probably already doing it. i made a skill scanner but it's also just easy to download a zip and inspect the contents... but people are loading these things remotely. i agree that it is easy to not install a pentester's magic skill, but the attack capabilities a skill can have are pretty insane. people should just make their own is my pov.