New todaygemini-4IMAGE

Gemini 4 Argon used after GPT and Claude block open-source vuln fix

Fixing a disclosed flaw in his own open-source project, the developer was refused by GPT under its cyber policy and downgraded by Opus 5.5, then continued with Gemini 4 Argon.

@ScarletKcArchived 7 hours agoOriginally posted 2026.10.29

Gemini 4 Argon used after GPT and Claude block open-source vuln fix (image 1 of 3)

Gemini 4 Argon used after GPT and Claude block open-source vuln fix (image 2 of 3)

Gemini 4 Argon used after GPT and Claude block open-source vuln fix (image 3 of 3)

KC @ScarletKcOriginally posted 2026.10.29

Open original on X

今天修 Hatoba 的一个安全漏洞，居然把 GPT 和 Claude 的安全防护机制都触发了。 GPT 直接以违反 Cyber Policy 为由拒绝回答，Claude Opus 5.5 更离谱，直接把整个会话标记成了 Cyber，强制降级到 Opus 4.8。最后只能换 Gemini 4 Argon 继续处理。问题是，这就是我自己开发的开源项目，漏洞也是公开披露的，我只是想让 AI 帮我排查和修复。现在连开发者修自己项目的安全漏洞，都得跟 AI 的安全机制斗智斗勇了吗？普通人想用 AI 修漏洞真的太难了。

This post quotes another post: x.com/scarletkc/status/2115757147820880025

x.com/scarletkc/status/2115765954000879891Archived 2026.10.30 09:59