[qsort]

It seems like they're doing something with the system prompt that I don't quite understand. I'm trying it in Claude Code and tool calls repeatedly show weird messages like "Not malware." Never seen anything like that with other Anthropic models.

[vessenes]

there's a line inside claude code mentioning to care about this. combined with new stronger instruction following behavior, you're going to be seeing it a lot unless you patch it out. or wait for a fix.