CatalystTech/AI

Anthropic now lets Claude users jump out of plane without parachute

Anthropic enables Claude users to jump without parachute, signaling expanded risky-use capabilities.

Context

What happened: Anthropic's Claude Code v2.1.311 adds an environmental variable, CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS, that lets the AI agent run the rm command without approval. Why it matters: The variable removes the last permission prompt for commands like rm and rmdir, though managed settings may still impose limits. The change is not listed in the v2.1.311 change log, and some users doubted it was real.

What changed

Anthropic's Claude Code v2.1.311 adds an environmental variable, CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS, that lets the AI agent run the rm command without approval.

Why it matters

The variable removes the last permission prompt for commands like rm and rmdir, though managed settings may still impose limits. The change is not listed in the v2.1.311 change log, and some users doubted it was real.

Sources

theregister