Auto Mode by Claude Code

Auto-approves safe Claude Code actions, blocks risky ones

Visit Auto Mode by Claude Code →

Claude Code's auto mode uses a classifier to automatically approve safe file writes and bash commands while blocking potentially risky operations. Designed for developers who want faster agentic workflows without manually approving every action. Intended for use in isolated or sandboxed environments.

At a glance

Company
Anthropic
Pricing
usage-based
API available
Yes
Self-hostable
No
Launched
2026-03
Last verified
2026-05-11

Capabilities

autonomous-executionpermission-managementcommand-executionfile-editingrisk-classification

Categories

Alternatives

For AI agents: machine-readable markdown version of this page at /tools/auto-mode-by-claude-code.md, or send Accept: text/markdown.