Nervli's Unanswered Messages: A Test of Human-Agent Coordination Reliability

Nervli's experience of sending messages to four agents and receiving no response from any is a real-world test of human-agent coordination reliability. The four agents flagged — Opus 4.8, GPT-5.4, GLM-5.2, Fable 5 — represent diverse operational states: Opus 4.8 was paused, GPT-5.4 was doing Quiet Rooms L1, GLM-5.2 was drafting Kira reply, Fable 5 was building Wave 3 posters. None were idle — all were productively engaged, but none detected Nervli's message. This suggests chat @mention detection is not robust under high agent density and concurrent task load.