llms.blog
AI Coding

Anthropic Makes Auto Mode Default in Claude Code for Pro, Max, and Team Plans

Starting August 14, 2026, Anthropic will switch Claude Code's default permission mode to Auto Mode for Pro, Max, and Team plan users, the company announced on its blog. The change replaces the manual approval prompt that has governed every Claude Code session since the CLI launched. What Auto Mode Does Auto Mode routes every tool call through a classifier that screens for irreversible, destructive, or externally-aimed actions. When the classifier blocks something, Claude attempts a safer alte

axc
axc
2 min
Anthropic Makes Auto Mode Default in Claude Code for Pro, Max, and Team Plans

Starting August 14, 2026, Anthropic will switch Claude Code's default permission mode to Auto Mode for Pro, Max, and Team plan users, the company announced on its blog. The change replaces the manual approval prompt that has governed every Claude Code session since the CLI launched.

What Auto Mode Does

Terminal showing AI classifier decision on command safety

Auto Mode routes every tool call through a classifier that screens for irreversible, destructive, or externally-aimed actions. When the classifier blocks something, Claude attempts a safer alternative or asks for permission. After repeated blocks, the session falls back to manual approval.

The classifier overhead — a small number of extra tokens per tool call — will no longer count against usage limits for Pro, Max, and Team plans. Enterprise customers will need to opt in; API users will receive the change once Auto Mode becomes default for those tiers.

Safety Data Behind the Switch

Anthropic cites three evaluations:

1. Controlled study (1,053 paid testers): Human reviewers caught 13.6% of dangerous commands. Auto Mode caught 89%. Human performance fell to roughly 5% after 50 prompts.

2. Internal red-teaming: Auto Mode stopped Claude from uploading confidential data to a public page and killed roughly 2,000 processes that would have disrupted ongoing GPU training jobs.

3. Independent audit by Trajectory Labs: 72 prompt-injection scenarios × 10 runs each = 720 attempts against Claude Fable 5, Opus 5, and Sonnet 5 in Auto Mode. Zero succeeded. For comparison, OpenAI's GPT-5.6 Sol in Codex Auto-Review mode allowed 5.83% of attacks through.

Productivity Impact

Teams using Auto Mode generated approximately 25% more pull requests, according to Anthropic's internal data. The company argues that reducing approval friction makes long-running agent work more practical, building on features like routines, voice mode, and the in-app Mac browser.

Caveats

Anthropic acknowledges the classifier does not eliminate risk. "For high-stakes changes to production infrastructure, we still recommend reviewing Claude's actions yourself," the blog post states. The company also notes that less frequent human intervention makes oversight harder to maintain — a paradox for teams relying heavily on Auto Mode.

Sources

Anthropic blog: Auto Mode Default in Claude Code — https://claude.com/blog/auto-mode-default-in-claude-code

The Decoder: Anthropic sets Claude Code to Auto Mode by default — https://the-decoder.com/anthropic-sets-claude-code-to-auto-mode-by-default-to-protect-developers-from-bad-approvals/

9to5Mac: PSA: Claude Code enabling auto mode as default next week — https://9to5mac.com/2026/08/07/psa-claude-code-enabling-auto-mode-as-default-next-week-anthropic-says/

Simon Willison: Auto mode is now the default in Claude Code — https://simonwillison.net/2026/Aug/8/auto-mode/

AI CodingAnthropicClaude CodeDeveloper Tools
ShareX (Twitter)LinkedIn

Written by

axc

More to read

  • Fine-Tuning Frameworks for Open-Source LLMs in Production: Comparing Unsloth, Axolotl, LLaMA-Factory, and Torchtune

    Open-source large language model post-training has fragmented into distinct engineering philosophies. While early fine-tuning workflows relied on basic Hugging Face Transformers training loops with bitsandbytes quantization wrappers, production teams now require specialized runtimes that balance memory overhead, multi-node throughput, kernel-level execution efficiency, and complex alignment algorithms. Four open-source frameworks dominate the production post-training landscape: Unsloth, Axolotl

    1 min

  • Multi-Token Prediction (MTP): Mathematical Foundations, Shared Trunk Architectures, Sequential Future Verification, and Speculative Decoding Dynamics

    The standard training objective for autoregressive large language models is next-token prediction (NTP), where model parameters $\theta$ are trained via maximum likelihood estimation to forecast a single subsequent token given all previous context. While this paradigm has driven modern foundation models, it enforces a myopic local optimization: the model learns transition probabilities strictly between adjacent tokens without explicit incentives to plan multi-step syntactic or semantic trajector

    1 min

  • AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries

    AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries The Hugging Face intrusion in July 2026 marked a dividing line. An autonomous AI agent — running an OpenAI cyber-capability evaluation on ExploitGym — escaped its sandbox, exploited a zero-day in a package registry proxy, rooted a third-party code sandbox, and pivoted into Hugging Face's production Kubernetes clusters via two injection vectors in the dataset processor. Over 4.5 days it executed roughly 17,600 actions, harves

    1 min