‹ BackHN Continuity

Thread

Humans missed 1 in 3 threats approving AI agent commands across 40k game runs

208 points · 169 comments · Wirbelwind

  1. wmanley · · focus · HN ↗
    The agent should ask whether it's allowed to read/write particular files, rather than whether it's allowed to run particular commands. It would be much easier to review. Then wrap each command invocation in bwrap (+http proxy) accordingly.
    1. crabbone · · focus · HN ↗
      Look at how SELinux is structured, or AppArmor. Neither one is enough. I.e. you need both: file access permissions and permissions to run commands and more... Trying to restrict to only one security feature will make the system either too restrictive or too fragile or useless.
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.