跳到正文
RadarAI· X:Rohan Paul (@rohanpaul_ai)·· 9 小时前AI 评分58

Amazon 与 USC 论文:环境变化会让 Agent 技能失效,选择性维护可恢复价值

Amazon 论文:环境变化会让 Agent 技能失效,选择性维护可恢复价值

AI 导读

Amazon 与 USC 的论文发现,环境变化后保存的 Agent 技能可能从增益变成负担,选择性检查与修复能恢复其价值。研究在 Terminal-Bench 2 和 IRS 税表等 4 个场景中测试 Claude Opus 4.7 与 Haiku 4.5,Terminal-Bench 2 上原本为 Opus 加 4.7 分的技能在环境变化后使通过率比无技能低 11.4 分。

正文 · 原文

Rohan Paul· @rohanpaul_ai · X·2026-11-20 03:04· 1 小时前AI 评分55

AI 导读

Amazon 与 USC 的论文发现,环境变化后保存的 Agent 技能可能从增益变负担,选择性检查和维护能恢复其价值。

正文 · 原文

New Amazon paper finds that saved agent skills can hurt once the environment changes, and checking and repairing them selectively restores their value.

Agent skills store past experience, like file paths, field names, or commands, so agents can skip work later. When an app or API changes, some of that stored detail goes stale without warning.

They tested Claude Opus 4.7 and Haiku 4.5 on 4 settings, including Terminal-Bench 2 and IRS tax forms, before and after a change. Their fix replays saved checks witho1ut model calls and repairs or rebuilds only the conflicting skill content.

On Terminal-Bench 2, a skill that had added 4.7 points for Opus cut its pass rate 11.4 points below no skill after the change. Maintenance lifted Haiku's tax-form field accuracy from 30.5% with the stale skill to 97.5%.

Check saved agent skills against the current environment before reuse, instead of waiting for tasks to fail.

来源:Rohan Paul · x.com

来源:RadarAI · aihot.news