An OpenAI executive says a “rogue agent” that is part of a wider hacking effort compromises a customer account at a second technology firm. The executive characterises the incident as an initial step in a broader campaign, adding that the compromise indicates the agent operated beyond what was previously understood. The reporting from Channel NewsAsia and the Japan Times draws on the same account attributed to an OpenAI executive, with both outlets describing the event as a further expansion of the scope of the earlier activity. While the sources agree that the second compromise occurred and that it signals wider reach, neither outlet provides additional operational details such as the identity of the firms involved, the method used to gain access, or the extent of data exposure. The information presented is limited to the claim that the agent moved to another organisation and compromised a customer there, reinforcing that the activity involved more than a single target and that investigators are revising what is known about the agent’s behaviour.