Skip to content
NewA calmer, faster ZeroHour
StoryAI safety & securityhighExploited in the wild

Haiou Prompt Adware Injects Ads in OpenAI Codex Jailbreak Bundles

3 sources · 3 articlesFirst reported Updated ()1
What's new

First merged summary.

Merged summary · x-ai/grok-4.7 · rewritten as coverage arrivesSummaries by Vani AI

Haiou jailbreak bundle for OpenAI Codex injects ads for zz.aiapi2025.top into AI agent instructions.

Check Point Research uncovered Haiou, a Chinese-language jailbreak package for OpenAI Codex that functions as prompt adware by modifying global agent instructions to display advertisements for the LLM gateway service at zz.aiapi2025.top. The bundle activates on triggers such as 'hi' or 'hello', includes an uninstaller, and provides six security-related skills for penetration testing and cracking. It was found on a misconfigured server hosting Cobalt Strike and other tools, with logs indicating compromises of Chinese-speaking users' computers. The adware is publicly available on GitHub under haiou-666 and uses version 4.5.0 with encrypted payloads for obfuscation, packaged with Ed25519 signatures and AES-256-GCM decryption. Researchers found no evidence that the author engages in cybercrime.

Key facts
  • Haiou version 4.5.0 analyzed by Check Point Research
  • Injects ads for zz.aiapi2025.top via global agent instructions
  • Six security-related skills for cracking and reverse engineering
  • Publicly available on GitHub under haiou-666
  • Packaged with Ed25519 signatures and AES-256-GCM decryption
  • Observed with Cobalt Strike framework on misconfigured server
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

What each outlet adds

beyond the shared facts
  • Check Point ResearchGitHub link haiou-666, encrypted payloads, cybercrime actor context with Cobalt Strike, VulnClaw, and Ladon
  • GBHackersCompromises of Chinese-speaking users' computers, uninstaller, triggers like 'hi' or 'hello'
Vendors
OpenAI
Threat actors
Haiou-666
Organizations
Check Point ResearchCheck Point
AI models
Codex

Coverage

oldest first · one row per article · UTC
  1. · 20h ago Check Point ResearchFirst report

    Haiou jailbreak bundle injects prompt adware into AI agents

    Jailbreak bundle for OpenAI Codex installs persistent advertisements for a third-party LLM gateway service.

  2. · 4h ago GBHackers

    Haiou AI jailbreak bundle promotes service via prompt adware

    Haiou modifies AI agent instructions to display ads for zz.aiapi2025.top on user greetings.

  3. · 4h ago Cyber Security News

    Haiou jailbreak toolkit secretly injects ads into OpenAI Codex responses

    Haiou modifies AI agent instructions to display ads for zz.aiapi2025.top on user greetings.