constraint compliance decay

Agent Memory Products Remember Your Conversation. Nobody Is Remembering Your Codebase

Every memory tool shipped this year stores what you said. Measured sessions of Claude Code auto compaction reduced 132,000 tokens of accumulated state to roughly 2,300, a 98% reduction, and what survives is the transcript rather than the understanding. Multiply that by 40 engineers separately paying a model to work out what the same billing service does, and the cost of not writing the answer down becomes visible.

Agent Skills Tell the Model How to Work. Nothing in That Loop Checks the Result

Skills are the best thing to happen to agent workflows in the last year, and we write them ourselves. This is not an argument against using them. It is an argument about what category of thing a skill is, because a large number of teams are now treating a code review skill as the answer to whether AI generated code is safe to merge, and a skill structurally cannot be that.

Your CLAUDE.md Was Measured, and It Was Roughly Net Negative

An ETH Zurich evaluation found machine generated context files lowered success rates and raised costs by roughly 20%, human written files bought about 4% on the harder benchmark, and Claude Code was the only agent tested where even developer written files failed to beat having no file at all. The most revealing result came from a control: strip the existing READMEs out and context files finally start helping.

Context Rot Is Not a Session Hygiene Problem. It Is an Architecture Problem

Compliance with an agent's own explicit constraints falls from 73% at turn 5 to 33% at turn 16 across 4,416 measured trials, and Claude Code auto compaction reduces 132,000 tokens of session state to roughly 2,300. Shorter sessions and cleaner instruction files treat the symptom. The window fills with low value tokens because the agent rebuilds an understanding of your codebase from raw source on every task.