Fight 'context rot' past 20–30 turns with periodic in-transcript reminders, hierarchical summarization, and judge agents

TechAhead·medium signal

Context rot is the measurable decay in an agent's ability to retrieve and act on information that is still technically in its window — coherence degrades beyond 20–30 turns, and most production agents break before 130K tokens despite 200K windows. One analysis attributed ~65% of agent failures to context drift or memory loss during multi-step reasoning, not model incapability. The cheapest non-obvious lever is re-injecting periodic reminders of the goal and key facts throughout a long transcript, layered with hierarchical summarization, offloading to external notes, and judge agents as runs scale.