Event ·Multi-source

Anthropic 联合 Gray Swan 举办 prompt injection 攻防竞赛

Anthropic 安全负责人 announced 与安全公司 Gray Swan 合作举办 prompt injection 竞赛，测试对象为 Claude 系列模型。Anthropic 表示过去几年在 Claude 模型的 prompt injection 防御上已有明显进展。随着 agent 承担更多任务，Anthropic 认为解决该问题已是全行业的紧迫事项。竞赛具体规则、奖金和时间安排待 Gray Swan 后续公布。

Current read

Anthropic 拉上 Gray Swan 办 Claude 的 prompt injection 攻防赛，搞安全的可以报名试试能不能攻破 agent。

11 sources66° AI heatLast updated 2026/10/15 17:55:32

Evidence

11 sources

Primary sourceBoris Cherny

2026/10/15 17:55:32View source

Related source 1Gergely Orosz

2026/10/15 16:50:26View source

Related source 2IT之家

2026/10/13 23:07:38View source

Related source 3Tech in Asia

2026/10/14 08:30:10View source

Related source 4Lxfater

2026/10/14 10:48:12View source

Related source 5Claude Code: GitHub Releases

2026/10/14 18:24:25View source

Related source 6HIT Consultant

2026/10/15 10:56:00View source

Related source 7Anthropic

2026/10/15 13:14:22View source

Related source 8Andrew Curran

2026/10/15 17:36:32View source

Related source 9宝玉

2026/10/15 23:29:34View source

Related source 10Nikkei Asia

2026/10/16 11:54:17View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story