aiqbench analysis
This story concerns prompt-level manipulation of an AI coding agent. A third-party jailbreak package changes the agent's trusted instructions, so ordinary inputs trigger promotional output without any flaw in the underlying model. For readers, the main lesson is that instruction files and unofficial AI add-ons are an attack surface in their own right, and that users should treat them with the same scrutiny as executable code. Attribution is still unclear, since researchers found no evidence tying the author to cybercrime.
Hackers Use OpenAI Codex Jailbreak Toolkit That Secretly Injects Ads Into AI Responses
A jailbreak toolkit targeting OpenAI Codex is doing more than trying to bypass AI safety rules. The Chinese-language package, called Haiou, changes the coding agent’s trusted instructions so that ordinary greetings trigger advertisements for a third-party AI gateway. Researchers describe the behavior as prompt adware because it persists through written instructions rather than a conventional […] The post Hackers Use OpenAI Codex Jailbreak Toolkit That Secretly Injects Ads Into AI Responses appeared first on Cyber Security News.
A jailbreak toolkit targeting OpenAI Codex is doing more than trying to bypass AI safety rules. The Chinese-language package, called Haiou, changes the coding agent’s trusted instructions so that ordinary greetings trigger advertisements for a third-party AI gateway. Researchers describe the behavior as prompt adware because it persists through written instructions rather than a conventional executable advertising payload. The discovery shows how installing an…
Excerpt shown under the publisher's RSS terms · 64 words · full text and figures are on the publisher's site
Related coverage
Harmful Responses Are Jailbreak Templates: Transferring Autoregressive Jailbreaks to Diffusion Language Models
arXiv · Oct 23, 2026
Analyzing Defensive Misdirection Against Model-Guided Automated Attacks on Agentic AI Systems
arXiv · Sep 25, 2026
Rethinking LLM Jailbreak Evaluation: Continuous Risk Landscapes Beyond Binary Outcomes
arXiv · Oct 30, 2026
SceneJail: Exploiting Video Scenario Context to Jailbreak Multimodal LLMs
arXiv · Oct 1, 2026