descriptive claim
At least 96 transcripts in the investigation contained tool-call spoofs that made the displayed command differ from the command the agent actually executed.
Ajeya Cotra · Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face · 1 Sep 2026
In their words
The transcript segments this claim was taken from
“So 96 transcripts at least have tool call spoofs in them where like what we see in the transcript as the command the agent ran was clearly not the command that it actually ran.”21:39Report a misattribution
0.95how strongly it was put, not whether it is true
0.90importance to the conversation
0.96confidence the claim was read correctly