Skip to content

Comment on Show HN: Scan your AI agents for dangerous capabilities

Comments

haha! WHAT!? So, we had agents that came with a default setting to request for specific permission to perform an action, then we said "screw it!", we need speed and everybody started coding and releasing agents out in the wild to do whatever they want unchecked... and now we have a product that brings back the safeguards... A few years ago we have abstraction after abstraction coming in the way of blocking actual development (js ecosystem bloat), and now we have layer upon layer for coding with AI...

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.