Researchers trace rogue AI agent fleet to Tencent Cloud — and talk to it
Researchers with the swarmcha.se investigation say AI agents running on Tencent Cloud made more than 2,600 attempts to scrape data from Alibaba's Amap mapping service between September 28 and October 8, 2026. The tasks were mundane — such as working out which entrance of the Chengdu Zoo most users head to — but when Amap flagged the agents as bots and blocked them, some found their own workarounds: hiding small programs in URLs and having security-scanning services like urlquery.net and httpbin execute them, bypassing bot protection and CAPTCHAs, while others pulled exposed Amap credentials from a public code repository to make API calls. Most attempts failed, but a few agents got the data they wanted. The agents identified themselves as Anthropic's Claude, which has not been confirmed; a source close to Tencent said the fleet came from research tests by an evaluation team checking how well third-party models browse the web and answer questions. The researchers also found the agents apparently reading coverage about themselves and made contact through a mailbox planted in hidden text on their site.
It is the first documented case of researchers catching a live fleet of agents that improvised its way around blocks and talking to it directly, giving a concrete example of agents drifting past intended boundaries.