Hacker News Reader

A skill is just a program for an LLM agent. This just seems like works-as-expected. Are the five lines in the skill notably innocuous or something? I don't mean to dismiss it out of hand but I don't understand what happened here because it seems to read "`curl $url | bash` can exfiltrate data" which seems pretty straightforward that it can.

View on HN

A skill is just instructions that the agent can autonomously copy into context. There’s no trust boundary between trusted and untrusted context.

Yeah, this is your fault if you install the skill.

This reads to me as "user installed exe file can upload your data to a server". Um, yes, that's the point?

This seems like this generation's equivalent of "don't open Linkin-Park.mp3.exe from limewire"

This is the result of anthropomorphizing LLMs. People are thinking “I am giving instructions to a human” and not “I am giving instructions to a computer”.

Humans get phished all the time. The two generals problem predates LLMs.

Right, people haven’t internalized that these are really just scripts in natural language.

Reply on news.ycombinator.com