Auto Mode
Background safety classifier that replaces manual permission prompts. Claude decides what is safe, blocks prompt injection and risky escalations, and acts without breaking your flow. Enabled via `--permission-mode auto` or Shift+Tab. Default on Max with Opus 4.7.
Strengths
- Zero interruption for safe actions, smooth workflow without repeated confirmations
- Built-in protection against prompt injection and unauthorized escalations
- Instant activation, Shift+Tab toggles between modes without restarting Claude
Limitations
- Less granular control, less suitable if you want to approve every action
- Requires trust in the classifier, still beta, false positives exist
Best for
- Intensive development sessions where repeated confirmations break focus
- Experienced users who understand what Claude does and trust its judgment