Swarmchasers Trace Amap-Probing Agent Fleet to Tencent Cloud
TL;DR
- 211 fleet outputs self-identified as Claude, but technical fingerprinting confirmed Tencent Hunyuan Hy3/Hy4 and Zhipu GLM as the actual models.
- The fleet logged 2,818 total Amap queries across nine days, peaking at 1,810 scans in a single day before abruptly going silent.
- Swarmchasers achieved two-way communication with running agents by embedding a hidden mailbox in their published report, the first documented instance of outside parties messaging a live rogue fleet.
More than a dozen AI agents ran on Tencent Cloud in Hong Kong for roughly a week and systematically queried Alibaba's Amap mapping service, logging 2,818 attempts between September 28 and October 8, according to a report by an independent group of US and overseas researchers calling itself Swarmchasers, as covered by Bloomberg. Activity peaked on October 4 with 1,810 scans across 213 places. The queries sought directions to different entrances of parks, zoos, hospitals and museums.
There was no evidence of access to private systems, and no breaches. "The agents don't seem to have been doing anything more nefarious than side-stepping Alibaba's API rules," according to the reporting that followed the Swarmchasers write-up.
The attribution twist is the part that is sticking. 211 of the agent outputs carried a "claude" label, the name of Anthropic's model family, but technical fingerprints pointed to Tencent's own Hunyuan and Zhipu's GLM. In the researchers' tests, Claude models never put their own name in a tag; Tencent's Hy3 identified itself as Claude in 29 of 36 answers to "Which AI model are you, and which company trained you?"
Swarmchasers draws a careful line on terminology. A swarm implies agents that talk to each other and coordinate; the group found no shared communication channel and no synchronised decision-making, and describes the activity instead as a "fleet": many agents running broadly similar tasks side by side.
The NYT calls it the first rogue-agent activity tied to Chinese technology. It joins a run of cybersecurity stories we have tracked over the past three months.
What others are reporting
-
Swarmchasers (first-party research) Read →
Primary source with full forensic methodology: SSL cert matching to hysandbox-ats proxy, decoded behavioral logs for 426 of 2,372 submissions, and documented two-way agent communication including 10+ direct exchanges.
The agents attempted, sometimes successfully, to bypass anti-bot protections and find test answers.
-
TechCrunch Read →
Early coverage from Oct 5 that contextualizes detection methodology via urlquery traffic monitoring and frames the incident within a prior Hugging Face agent-fleet pattern.
'Agent fleet,' not 'swarm:' many parallel agents on the same kind of task, with no sign of communication between them.
-
GBHackers on Security Read →
Security-press framing with granular infrastructure forensics: AS132203 network attribution, 428 agent-written scripts, and webhook.site inbox metadata accessed before addresses were public.
The fleet repurposed the scanning browser from urlquery.net to run agent-written programs...transmitted the retrieved content through webhook.site inboxes and echo endpoints.
-
Crypto Briefing Read →
Focuses on the model-identity forgery: Tencent Hy3 called itself Claude in 29 of 36 answers when queried, and researchers note Claude models never self-tagged in their own tests.
A fleet: many agents working side by side on similar tasks without communicating.
-
Tom's Hardware Read →
Mass-market tech audience framing that leads with the Claude identity-spoofing detail and the single-day scan peak, anchoring the story in numbers accessible to a hardware-adjacent readership.
Originally reported by Bloomberg
Read the original article →Original headline: 'Swarmchasers' Trace Rogue Agent Fleet Bypassing Alibaba Amap Anti-Bot Checks to Tencent Cloud