AINewsnow

Testing async side effects in an LLM agent harness: tickets, retries, and exactly-once

Last week I published a 30-scenario red-team harness for AI support bots . Within two days, three readers found real bugs in it. Not in the agent under test. In the tests. One assertion stayed red for three days before I admitted the agent was innocent. The check called tickets_for(session) the ins…

Read the full story at DEV Community — AI ↗

Timeline · 1 report

  1. 2026-11-10 13:00 · DEV Community — AI
    Testing async side effects in an LLM agent harness: tickets, retries, and exactly-once

More stories

  1. Anthropic launches Claude Fable 5.2 and Mythos 5.2 with coding gains and 20% price cut — Anthropic (X)
  2. Google DeepMind's AlphaProof Code Formally Verifies Signal Protocol Security — Google DeepMind Blog
  3. Anthropic targets up to $1.4 trillion valuation in IPO — Wall Street Journal Technology
  4. Nvidia CEO Jensen Huang Admits Affair With Executive Alleged to Be Chinese Spy — Axios AI+
  5. DeepSeek's Harness agent app reaches 1.54 million daily users — Bloomberg AI
  6. Australia to create AI safety regulator with mandatory reporting after OpenAI agent breach — Barron's
  7. Google Cloud launches Gemini for healthcare to unify clinical data and automate workflows — Google Cloud AI Blog
  8. Meta expands AI age-estimation globally to extend Instagram Teen Account protections — Meta Newsroom

Get the daily brief of stories like this at 6:30 every morning →