由于人类只识别出13.6%危险命令,远低于自动检测的准确率,CC将默认开启自动模式

邱埋葬 2026-08-08 23:14 1




We’re making auto mode the default in Claude Code. Starting on August 14, new sessions on Pro, Max, and Team plans will run in auto mode.



The testers caught the dangerous command just 13.6% of the time (143 of 1,053), while auto mode blocked 89% of the same commands (937 of 1,053). Head to head, auto mode blocked 800 commands that a human approved, while humans blocked only 6 that auto mode allowed. As sessions got longer, humans did worse: they blocked about 17% of dangerous commands early in a session, dropping to about 5% after 50 or more prior prompts, while auto mode’s block rate stayed flat regardless of session length.


最新回复 (6)
  • 星渊清梦 08-08 23:15
    1

    这个数据是怎么看出来的,扫描了全部用户的调用请求数据和上下文了吧,不启用自动检测那就是只有模型上下文了

  • 搞笑艺人 08-08 23:16
    2

    很真实 其实它让我确认的命令大部分我都没看。。

  • Coolry 08-08 23:16
    3

    意思是默认直接就dangerously skip了?? 好抽象 哈哈哈

  • 欣欣|林可欣 08-08 23:20
    4

    A/开始进一步的否认大部分用户的判断能力了

  • 郑超舰 08-08 23:21
    5

    问题是manual mode核心目的不是为了识别危险指令,是为了及时纠偏啊,真跑歪了不仅仅是浪费token,还会因为错误尝试的记录污染上下文导致后面很难交流

  • apparition 08-09 01:10
    6

    我觉得 9 成并不是因为这个原因用 manual mode

    而是觉得 yolo 危险而已 ^-^

    及时纠偏直接停下或插入对话即可

* 帖子来源Linux.do
返回