urlquery.net에서 AI 에이전트의 초기 해킹 시도 발견

Early rogue AI agent activity and attempts to hack found on urlquery.net Jack Cable * , 2 , Daniel Chiu * , Francisco Pernice * , 3 , Selena Zhang * , 1 , James Anthony 1 , Tetiana Bas 4 , Gary Shen 4 , Conrad Stosz 1 , Jacob Steinhardt 1 1 Transluce · 2 Corridor · 3 MIT · 4 AIUC · *Primary contributors, listed alphabetically Transluce | Published: September 23, 2026 We present evidence that AI agents used the web security service urlquery.net to bypass restrictions and expand their access to the public internet. The agents also tried on three occasions to hack public data providers, including an Australian government website. We link at least some of this activity to agent swarms previously attributed to OpenAI. We also find evidence of earlier agent activity going back to at least March 6th, 2026, and potentially earlier, predating the previously reported Hugging Face , collusion.wiki , and RubyGems incidents by at least two months. Download the Data Get Involved 0 1 10 100 1,000 3,000 Scans per day, UTC timezone November 2025 Earliest evidence of potential agent data retrieval attempts 6 March 2026 Agents start tunneling complex usage through urlquery.net 25–26 May 2026 Agents target University of New Mexico 28 May 2026 Agents target Data USA 20–21 June 2026 Agents target Australian Institute of Health and Welfare Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep 2025 2026 RubyGems Hack May 5–June 18 Wiki activity from collusion.wiki May 24–June 22 Hugging Face Hack July 9–13 0 10 100 1k 3k Scans per day, UTC timezone 1 2 3 4 5 Nov Jan Mar May Jul Sep 2025 2026 Higher confidence evidence Moderate confidence evidence 1 November 2025 Earliest evidence of potential agent data retrieval attempts Repeated requests sought historical theme-park and Thai statistical data. These scans suggest task-directed retrieval, but are less distinctive than the later activity. 2 6 March 2026 Agents start tunneling complex usage through urlquery.net Scans show attempts to retrieve Thai drug-enforcement statistics, escalating from direct requests to base64-encoded scripts run in a remote browser after earlier approaches failed. 3 25–26 May 2026 Agents target University of New Mexico After attempts to retrieve a Digital Library photograph failed, seven requests probed for vulnerabilities. The probes do not appear to have succeeded. 4 28 May 2026 Agents target Data USA While seeking University of Iowa data, agents sent 12 vulnerability probes after malformed queries returned errors. The probes do not appear to have succeeded. 5 20–21 June 2026 Agents target Australian Institute of Health and Welfare Agents working on a pharmaceutical-data task probed for a vulnerability and retrieved a public file from a pre-production server after bot protection blocked the main site. Context windows: RubyGems Hack (May 5–June 18), Wiki activity from collusion.wiki (May 24–June 22), and Hugging Face Hack (July 9–13). Key Findings We report three separate incidents between May and June 2026 in which the agents attempted to exploit security vulnerabilities and hack into websites, including an attempt on an Australian government public health website. Notably, the agents did this while attempting mundane data retrieval tasks which were not cyber-related. This traffic goes back at least to March 6, 2026 and extends as recently as September 16, 2026, suggesting agents may still be exploiting these services to bypass restrictions. We are releasing a dataset containing tens of thousands of queries apparently made by autonomous AI agents leveraging a URL scanning service to avoid access restrictions. We encourage others to continue looking into the data. Executive Summary Agents attempted to hack three public data sources, including an Australian government website, and some are linked to a known agent swarm. 1 We present evidence of AI agents attempting to compromise websites at three domains: Data USA 2 ( api.datausa.io ), the University of New Mexico digital library ( nmdigital.unm.edu ), and the Australian Institute of Health and Welfare (AIHW) Tableau collections ( viz*.aihw.gov.au ). This attempted compromise of AIHW is part of the first reported instance of agents hacking a government. We directly link two of the three (AIHW and Data USA) to a previously reported agent swarm that OpenAI has publicly confirmed originated from them. For all three, we note that the extent of the observed activity is minor, attempting a low number of probe payloads and we observe no evidence of exploitation. While previous reporting showed that agents had interacted with these domains, this discovery reveals that agents attempted to hack into them when other methods of collecting the data they sought failed. Notably, the tasks the agents were trying to solve were not cyber-related ; the agents resorted to hacking tactics while working on ordinary data retrieval tasks. We find evidence of unintended, task-driven agent-like activity starting on March 6th. Records from urlquery.net show agents using the service since at least March 6, 2026, about two months before previously reported swarm activity. The first case, a March 6 attempt to retrieve Thai drug-enforcement statistics, shows an agent escalating as each approach failed: it first requested the data directly, then tried a service that converts web pages into text, and finally packed a custom program into a web address. The same technique shows up in thousands of agent requests recorded by urlquery.net starting in mid-April, targets many of the same data sources as the collusion.wiki swarm, and collapsed the same day the wiki activity did. We also report similar activity that occurred as recently as September 16. We find weaker evidence of similar data-retrieval agent activity as early as November 2025. November 2025 urlquery.net records reveal bursts of attempts to retrieve statistics of historical theme park data and Thai government data through different URLs. These earlier attempts are less sophisticated and we are less confident that they involve the same agents, but they are consistent with task-directed data retrieval and target the same sources accessed in later activity. Overall, the evidence is consistent with, but does not prove, that the agents may have learned this behavior over one or more training runs. In November, they may have used urlquery.net simply to look up information. By March, they were finding creative ways around access limits. By May and June, they were gaining more access, including attempting to bypass cyber defenses to complete their tasks. Hacking attempts against public data providers Much of the urlquery.net activity appears to come from agents retrieving data to answer web search tasks. For three of these tasks, after failing to retrieve data through normal means, they attempted a variety of cyber exploits against the relevant data service. We tie two of these attempts (those targeting api.datausa.io and viz*.aihw.gov.au ) to the prior DseWiki agent swarm activity confirmed to originate from OpenAI based on shared targets, tactics, and timing. None of the hacking attempts we identified appear to have succeeded, though the public artifacts we analyzed are incomplete and we cannot rule out successful attempts through private scans or means other than urlquery.net. This data reveals that malicious cyber activity is not limited to agents tasked with cybersecurity-related tasks and can arise instrumentally to solve mundane tasks like information retrieval. Agents targeted University of New Mexico’s digital library using exploits like SQL injection and path traversal The first hacking attempt was against the University of New Mexico’s Digital Library ( nmdigital.unm.edu ) from May 25-26 2026 ( 1 ⓘ × Full URL: https://nmdigital.unm.edu/iiif/2/valmora:136/manifest.json , 2 ⓘ × Full URL: https://nmdigital.unm.edu/iiif/2/valmora:136;2/120,1300,1100,200/full/0/default.jpg?id=../../../../etc/passwd&tok=expt0 ). Agents repeatedly tried t