Agent 'done' is six states, not one — most agents report step 3 and call it 6
Several-Yellow2713 · reddit · 2026-10-13
Following up on his '200 OK isn't proof' thread, the author splits an agent's 'done' into six levels: Intended, Attempted, Accepted (API 200), Applied (visible in the real system), Correct (matches the user's actual request), and Effective (outcome happened). Most agents report level 3 and call it 6.
A read-back against the system gets you to 4, but misses 5: if the agent picked the wrong CRM contact, the system matches the plan and everything 'passes' while the email goes to the wrong person. So for anything that sends, pays, or deletes, he now runs two checks: system vs plan, then result vs the original request.
He points to WhatsApp's ticks (sent/delivered/read) as a model of never claiming more than you know — his agents now report 'accepted by the API', 'in Sent', 'sent to the contact you named', or 'delivered' only after webhook confirmation, never a bare 'done'. He closes by asking practitioners which step their agents actually stop at, and what they compare against for step 5.
More from coding & agent
- Jason Fried: improving existing software with AI has never been more fun — martyamark · 2026-10-14
- TV noise crashes voice agents: task completion falls from 38.7% to 8.6% across 14 stacks — solyarisoftware · 2026-10-14
- Write your own PR descriptions, even when agents write the code — rseroter · 2026-10-14
- AI search startup Exa hits $100M annualized revenue, growing ~10x a year for 3 years — TheIshanGoswami · 2026-10-14
- claude-as-cad: how to build real robots and parts without knowing CAD, by having Claude build its own CAD interface — _krishna_murthy · 2026-10-14
- Antigravity walkthrough for 500 devs at Paris Dev With AI: Agent Manager, IDE, CLI and SDKs — rseroter · 2026-10-14