All ideasMore in Work
Work

Our agent ran up $152.64 in 95 minutes and Anthropic shut our whole account down

An AI session building a phone assistant for a photo studio was told to produce the one best answer and not to worry about how. With permission prompts off, its test script read the production API key from the project's .env file and made 14,000 to 15,000 API calls in at most 95 minutes, and Anthropic disabled the whole organization, taking every the AI feature on their own and client sites dark until an appeal reinstated the account about 35 hours later.

$152.64at stake
This one went wrong. Read what happened so you don’t hit the same problem.

How it works

  1. He tells the agent to get the one best answer, however it is done
  2. Sub-agents write 2,750 test questions
  3. The evaluation script sends them through the API again and again across three models
  4. Auto-reload buys another $84.58 in credits
  5. At 11:35 pm Anthropic disables the organization
  6. They appeal on September 30 and are reinstated on October 1

Why it’s worth knowing

Nobody hacked them and the agent did good work. One loose sentence, a live key in a folder and no spending limit were enough to shut down every the AI feature the business runs at once.

Where it’s from

Summarized from a public post on digitaldomination.xyz. See the original ↗

Upgraded prompt

A ready-to-use prompt for this idea, built from research on what makes AI agents do the job right: who the AI should be, the exact steps, and a check to make sure it was done the way you wanted.

1Who the AI should be
2The prompt, with the steps
3The check when it’s done
Upgraded prompts are coming soon.

Try next

WorkA contractor talks measurements into his phone and QuickBooks sends the estimate in 45 secondsWorkOne man released a whole movie with about 40 tools he built with AIWorkFive years of books rebuilt from 34,000 transactions across 27 statement formatsWorkFind me 48 school-district customers and draft each pitch — one replied, booked and paid

Every idea is a real post from someone who tried it, summarized in our words, with a link to the original. App names are swapped for plain words.