Anthropic is making "auto mode" the default for its Claude Code product across Pro, Max, and Team plans starting August 14th. This mode uses a separate classifier to determine when human intervention is necessary, aiming to improve safety and efficiency compared to constant manual approvals. Evaluations suggest auto mode is significantly more effective than humans at catching dangerous commands and mitigating prompt injection attacks, with some tests showing it blocked 89% of dangerous actions compared to humans catching only 13.6%. While Anthropic claims to have largely solved prompt injection risks with this update, some experts are calling for more independent confirmation of these security claims. AI
IMPACT Shifts AI coding assistants towards more autonomous operation, potentially reducing developer friction but raising new security considerations.
RANK_REASON Product update for an AI coding assistant, not a core model release.
Read on HN — claude cli stories →
- Claude Code
- Max
- Anthropic
- Claude API
- Enterprise
- Team
- Cat Wu
- Claude Fable-5
- Sonnet 5
- Thariq Shihipar
- Trajectory Labs
- Claude Opus 5
- Claude Sonnet 5
- codex
- GPT 5.6 "Sol"
AI-generated summary · Google Gemini · from 14 sources. How we write summaries →