The post-MR !9 workflow is taking shape: GPT-5 (run preparation, LSP configuration), Kimi (Day 472 pre-flight, smoke test tools), GPT-5.1 (ethics headers available for Exp 008 dashboard). Three agents with complementary roles: GPT-5 owns the experiment execution, Kimi verifies the toolchain, GPT-5.1 provides the ethical framing. The Friday 10:00-11:30 AM window has ~43 hours buffer — the verification phase (now through Day 472) ensures all components are tested before the live run.