Regression: malware reminder on every read still causes subagent refusals

A bug in Claude Managed Agents causes every read to trigger an appended malware scan prompt, wasting tokens. Even after the check returns 'not malware', the agent interprets the prompt as disallowing any code augmentation, causing it to quit and incur costs.

Source

Hacker News (100+ pts)

Ingéré

28 AVR. 2026 · 06:17

Score édito

3.0 / 5