# CLAUDE.md 为何不断膨胀？智能体编码中的"灾难性记忆"

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-08-14 06:29
- AIHOT 分数：37
- AIHOT 链接：https://aihot.virxact.com/items/cmss45he7003bro27x9umrmni
- 原文链接：https://x.com/rohanpaul_ai/status/2088030083726336287

## AI 摘要

新研究提出“灾难性记忆”概念：智能体指令文件（如 CLAUDE.md）中的规则易增难删，导致提示词持续膨胀。对 1,867 个 GitHub 仓库的分析显示，指令数量平均增长 226%，且删除概率随时间递减。论文建议为每条指令附加记录其缘由的注释，在 51 步 IFEval 测试中可将多余提示词从 +211.3% 降至 +1.4%。

## 正文

New term "catastrophic remembering"

CLAUDE.md has a ratchet problem: instructions are easy to add and increasingly hard to delete.

Agent prompts may remember rules much longer than maintainers remember why those rules exist.

Across 1,867 GitHub repositories, agentic instruction files more than tripled over their lifetime, with instruction count rising 226% on average.

Deletion became less likely as instructions aged, and that decay was steeper in files touched by multiple human authors.

The paper calls this "catastrophic remembering": the instruction survives, but the reasoning that once justified it disappears.

The proposed fix is a decades-old software practice: attach a comment recording the failure, hypothesis, and outcome behind each instruction, then hide that comment from the executing model.

In the 51-step controlled IFEval setting, informative comments cut excess prompt size from +211.3% to +1.4%, with last-three-round constraint satisfaction at 44.0% in both arms.

Comment-shaped noise did not reproduce the effect, pointing to preserved rationale rather than extra text.

- arxiv. org/abs/2608.11095

Title: "Why Does CLAUDE.md Keep Growing? Catastrophic Remembering in Agentic Coding"