Apollo Red-Teams Anthropic's Auto Mode, Cutting Classifier Miss Rate to 7%

MariusHobbhahn · x · 2026-08-12

Apollo Research conducted its first external red-teaming campaign for Anthropic’s auto mode in Claude Code.

After hardening the system based on Apollo’s findings, the classifier's miss rate dropped from 12% to 7%. Auto mode is now the default permissions mode in Claude Code.

Original post →

More from coding & agent

coding & agent channel →