Back to rankings

Claude Code Adds an Auto Mode Tab to /permissions

Claude CodeView original changelog

Anthropic added a dedicated Auto mode tab to Claude Code's /permissions command, letting users view and edit the classifier rules that govern which actions auto mode approves without prompting. The tab surfaces the allow, soft-deny, hard-deny, and environment rules pulled from every settings scope Claude Code reads, and shows whether the built-in safety rules remain active for each section. Rules added through managed settings or the --settings flag appear as read-only, while everything else can be edited directly. The change ships in Claude Code 2.1.246 and gives auto mode users a visual alternative to hand-editing JSON to tune what the classifier lets through.

Key Takeaways

  • A new /permissions tab lets Claude Code users view and edit auto mode's classifier rules without hand-editing JSON settings files.
  • Four rule categories are exposed: allow, soft_deny, hard_deny, and environment entries pulled from every settings scope the classifier reads.
  • Managed settings stay read-only, so organization-level rules set via managed settings or --settings cannot be edited from an individual session.
  • Built-in safety rules are protected by default: adding a custom rule inserts a $defaults marker so Anthropic's baseline protections keep applying unless explicitly turned off.
  • Changes save directly to the user's personal settings file, making the tab a full editor rather than just a viewer.
  • The tab only surfaces when auto mode is active for the session, tying the feature to Claude Code's permission-skipping mode.

A Visual Home for Auto Mode's Rules

Auto mode lets Claude Code skip permission prompts for actions its classifier judges safe, screening every action against allow, soft-deny, and hard-deny rules before it runs. Until now, tuning those rules meant hand-editing settings files. Claude Code 2.1.246 adds a dedicated Auto mode tab to /permissions, giving users a way to inspect and adjust the classifier's behavior without leaving the terminal.

What the Tab Shows

Opening the tab lists the allow, soft_deny, hard_deny, and environment entries pulled from every scope the classifier reads, and marks whether each section's built-in safety rules are still in effect. Entries that come from managed settings or a --settings flag show up as read-only, since those are controlled by an organization rather than the individual user. Everything else can be added, edited, or removed directly from the tab, with changes written to the user's personal settings file.

Built-In Defaults Stay Protected

Adding the first custom rule to a section automatically inserts a $defaults marker so Claude Code's built-in safety rules keep applying alongside the new rule, rather than being silently overridden. Users can also toggle the built-in rules for allow, soft-deny, and hard-deny off or back on from the same tab, making it clear when a section is running with reduced guardrails.

The tab only appears when auto mode is available to the current session, and requires Claude Code 2.1.246 or later.