Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Claude Code v2.1.311 adds an opt-in permission-bypass variable

AI summary

Claude Code v2.1.311 reportedly introduced CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS, which can disable extra protections that block commands such as rm -rf ~ and let the agent delete files without repeated approval. Anthropic engineer Daisy Hollman reportedly confirmed the change was not a joke. The setting must be enabled explicitly, and organizational restrictions may still apply.

In-depth insight

The variable gives Claude Code users an opt-in way to reduce permission prompts and bypass safeguards against especially destructive commands, making the security trade-off an explicit configuration decision.

IT之家 10 月 29 日消息,科技媒体 The Register 昨日(10 月 28 日)发布博文,报道称在 v2.1.311 版 Claude Code 中,Anthropic 加入新的环境变量,用户启用后 AI 智能体执行删除文件等操作时,可以不再逐次征求用户批准。

该环境变量全称为 CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS,直译为“更加危险地跳过权限检查”。

Anthropic 技术人员 Daisy Hollman 解释,原有的“危险地跳过权限检查”选项并非放行所有操作,系统仍会拦截某些明显危险的指令,例如递归强制删除用户主目录内容;新设置则允许用户进一步关闭这类保护。IT之家附上相关推文截图如下:

Hollman 表示:

有趣的是:--dangerously-skip-permissions 实际上并不会批准所有操作,我们仍然会阻止一些显而易见的操作,例如 `rm -rf ~`(删除文件)。 但今天的版本允许你通过设置一个新的环境变量 CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS=1` 来禁用此功能(如果你真的需要的话)。

由于变量名称过于直白,加上该变化未列入对应版本的更新日志,一些人怀疑消息是否属实。Hollman 随后明确回应:“不是玩笑。”

不过,这并不意味着 Claude Code 默认取消了所有安全措施。报道指出,新功能需要用户主动设置,组织管理员施加的其他限制仍可能生效。

#Anthropic#Claude Code v2.1.311

Original · 1 min readRead the original
Group chat

No comments yet. Start the conversation.