Codex

The council worked. One request went to Claude, Codex and Kimi, they answered independently, and nothing touched the lab without a human approval bound to a hash of the exact plan. It also lived entirely in a terminal, which meant that to approve a run I had to type a run id like hc-20260822-6fbfa5 and then re-type the original 350-character request verbatim, because resume rebuilds the context pack from it. Get one character wrong and you have quietly changed the inputs the agents were judged on. ...

I have three coding agents with access to this lab. Claude Code does most of the building. Codex reviews and runs the scans. Kimi audits. Until this week they had never exchanged a word — each got its own context, its own session, and its own chance to be confidently wrong with nobody checking. So I built a thing that makes them work one request together, and put a human approval gate in front of anything that changes the lab. ...

I run two coding agents against this lab. Codex found the problem and Claude Code fixed it, and keeping those two jobs in different hands turned out to be the most useful part of the whole exercise. The short version: a full vulnerability scan reported 2042 actionable findings — 46 critical, 764 high — and 1751 of them were in container images. Ten waves of updates later, the same scanner reports 1497 actionable, 21 critical, 570 high. ...