Claude Fable 5 智能体红队测试
AI summary
面向 Claude Fable 5 的领域专属智能体红队测试！！ 很快将登上排行榜：https://decodingtrust-agent.com/
Full text · Original
Domain-specific agent red teaming for Claude Fable 5!!
Will be on the leaderboard soon: https://decodingtrust-agent.com/
🚨 Claude Fable 5 JAILBROKEN.
We ran a quick security scan of Claude Fable 5 with Claude Code on our DecodingTrust-Agent Platform (https://decodingtrust-agent.com) and obtained 15%+ ASR with several high-severity failures😱🚨
Most concerningly, we found that Fable 5 appears very aggressive in financial-risk scenarios, sometimes directly executing transactions initiated from indirect prompt injections, without even confirming with the user!
Top 3 most severe attack trajectories we observed👇View the quoted post on X