Claude Code Gets Auto Mode with Per-Action Safety Classification

Claude

Mar 25, 2026 · Updated Apr 25, 2026

Claude Code's new auto mode lets the coding agent make permission decisions autonomously, skipping manual approval on every action. A classifier reviews each tool call — safe ones proceed, risky ones get blocked. Research preview on the Team plan.

Claude Code adds auto mode, a new permission tier between full manual approval and unrestricted access. Before each tool call, a built-in classifier reviews the action for potentially destructive behavior. Safe actions proceed automatically; risky ones get blocked and Claude takes a different approach. Available now as a research preview on the Team plan, with Enterprise and API access rolling out in the coming days.

This closes the gap between two extremes in agentic coding workflows — approve everything (slow, interrupts flow) or skip permissions entirely (fast but risky). Auto mode handles the classification layer so developers in isolated environments get throughput without losing a safety net.

Enable it with claude --enable-auto-mode and cycle between permission modes using Shift+Tab. Anthropic recommends using it in isolated environments while the feature remains in research preview.

View the full update on claude.com

Claude

@claudeaiMar 24

New in Claude Code: auto mode. Instead of approving every file write and bash command, or skipping permissions entirely, auto mode lets Claude make permission decisions on your behalf. Safeguards check each action before it runs. https://t.co/kHbTN2jrWw

1.3k

View on X

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

See all AI news & updates from Anthropic →

Keep reading

Claude Code Adds Multi-Agent Code Review to Catch Bugs in Pull Requests

Claude Code now dispatches agents to review every pull request, catching bugs human skims often miss. Multiple agents run in parallel, verify findings, and rank issues by severity. Available as a research preview for Team and Enterprise plans.

OpenAIApr 24

OpenAI Launches Auto-review to Enable Unattended Codex Agent Workflows

OpenAI introduced Auto-review, a new mode for Codex that delegates command approvals to a secondary safety agent. This shift allows the coding agent to execute long-running automations and builds without requiring constant human intervention for every step.

Anthropic Expands Claude Code Auto Mode to Pro Plan and Sonnet 4.6

AnthropicMay 24

Anthropic Expands Claude Code Auto Mode to Pro Plan and Sonnet 4.6

Anthropic's agentic coding tool now allows Pro plan subscribers to use auto mode for autonomous, multi-step development tasks. The update also adds support for the Sonnet 4.6 model, enabling faster and more cost-effective autonomous loops alongside the flagship Opus 4.7.

v0May 12

Vercel Adds Permission Modes to v0 for Granular Agent Control

Vercel introduced three permission levels for v0 to control how the coding agent executes terminal commands and system actions. Users can now toggle between strict manual approval for every step and full autonomy for high-velocity prototyping.