Jev adds hard approval locks to a 24/7 office agent
A Sonnet 5.5 agent handles email, invoices and reports while Jev blocks high-risk actions such as amounts over $500, contracts and prices for human review, letting the rest through automatically.
ANTHROPIC + OPENAI LEAKED A FILE THAT TAKES 7 HOURS OF ENGINEERING WORK OUT OF EVERY DAY FOR $6 A WEEK 45 hours of your time is $1,800, the tokens that eat them cost pocket change email → support → invoices → social → reports one agent on Sonnet 5.5 rolls across your week and paints 45 of the 50 tiles as done OpenAI Dots feeds it the inbox and the CRM, Jev locks the five hours the agent is never allowed to touch 36 of 40 emails never reach you, some closed with a reply, the rest just filed four need a living yes, and those four are the only ones worth anything Jev lets the safe ones through and turns the rest back: money over $500, contracts and prices always come back to a person that rule was written before the agent started, not after the first wrong charge went out the speed here comes from the locks, not the model, because without them 90% automation is a liability, not an asset and here is the brutal part of this file: a thousand actions a week at 98% right is twenty wrong ones 90% of the hours is not 90% of the value, and the five hours left decide whether the other 45 went right finding those twenty mistakes is the only job you have left save this and paste it into Claude Code - let it take those 45 hours off you ↓
This post quotes another post: x.com/sprytixl/status/2107763347097534680