“While most prompts are likely for safe, routine commands, an approval rate that high suggests many users are clicking through reflexively rather than reviewing each command”

The New Stack · Aug 7 · Anthropic makes Claude Code's Auto Mode the default, citing human approval fatigue

“the AISI testing parameters were not representative of any of our production models”

BBC · Aug 4 · Anthropic's Mythos AI created fake identities to hack developers, AISI reveals

“Cross-session messaging lets Claude deliver a message from one of your Claude Code sessions to another.”

Claude Code documentation · Anthropic launches cross-session messaging for Claude Code