Did a Chinese University Hacking Competition Target a Real Victim?

Capture the flag hacking contests at security conferences generally serve two purposes: to help participants develop and demonstrate computer hacking and security skills, and to assist employers and government agencies with discovering and recruiting new talent.

But one security conference in China may have taken its contest a step further—potentially using it as a secret espionage operation to get participants to collect intelligence from an unknown target.

According to two Western researchers who translated documentation for China’s Zhujian Cup, also known as the National Collegiate Cybersecurity Attack and Defense Competition, one part of the three-part competition, held last year for the first time, had a number of unusual characteristics that suggest its potentially secretive and unorthodox purpose.

Capture the flag (CTF) and other types of hacking competitions are generally hosted on closed networks or “cyber ranges”—dedicated infrastructure set up for the contest so that participants don’t risk disrupting real networks. These ranges provide a simulated environment that mimics real-world configurations, and participants are tasked with finding vulnerabilities in the systems, obtaining access to specific parts of the network, or capturing data.

There are two major companies in China that set up cyber ranges for competitions. The majority of the competitions give a shout out to the company that designed their range. Notably, Zhujian Cup didn’t mention any cyber range or cyber range provider in its documentation, leaving the researchers to wonder if this is because the contest was held in a real environment rather than a simulated one.

The competition also required students to sign a document agreeing to several unusual terms. They were prohibited from discussing the nature of the tasks they were asked to do in the competition with anyone; they had to agree not to destroy or disrupt the targeted system; and at the end of the competition, they had to delete any backdoors they planted on the system and any data they acquired from it. And unlike other competitions in China the researchers examined, participants in this portion of the Zhujian Cup were prohibited from publishing social media posts revealing the nature of the competition or the tasks they performed as part of it.

Participants also were prohibited from copying any data, documents, or printed materials that were part of the competition; disclosing information about vulnerabilities they found; or exploiting those vulnerabilities for personal purposes. If a leak of any of this data or material occurred and caused harm to the contest organizers or to China, according to the pledge that participants signed, they could be held legally responsible.

“I promise that if any information disclosure incident (or case) occurs due to personal reasons, causing loss or harm to the organizer and the country, I, as an individual, will bear legal responsibility in accordance with the relevant laws and regulations,” the pledge states.

The contest was hosted last December by Northwestern Polytechnical University, a science and engineering university in Xi’an, Shaanxi, that is affiliated with China’s Ministry of Industry and Information Technology and also holds a top-secret clearance to conduct work for the Chinese government and military. The university is overseen by China’s People’s Liberation Army.

Neither the university nor several cosponsors of the contest responded to WIRED’s inquiry about the competition.

Dakota Cary, a strategic advisory consultant at security firm Sentinel One, and Eugenio Benincasa, senior cyber defense researcher at the Center for Security Studies at ETH Zurich university in Switzerland, discovered the unusual contest and its requirements while researching China’s hacking competitions. The two have written a report for the Atlantic Council and plan to present their findings on Thursday at the Labscon security conference in Arizona.

They acknowledge that there is no hard evidence that the competition was used to attack a real-world target and that the evidence they do have is circumstantial. But they said they are 85 percent confident that this is what occurred because no alternative explanations make sense.

“There are a lot of good alternative explanations, and none of them have supporting evidence,” says Cary, who is fluent in Mandarin and has studied China’s offensive hacking capabilities extensively for years.

One possible alternative is that the target of the attack was a domestic company or organization that cooperated with the contest in order to give students experience in finding vulnerabilities in a real network and also provide the company with a threat assessment that could help it better secure its network against real-world adversaries.

Cary, however, says such exercises in China are called “crowd-testing” contests. But nowhere in the description of the Zhujian Cup does this phrase appear. “If it’s a real CTF exercise, why are [participants] deleting data and … backdoors?” he asks. And why are students told they could be held legally responsible if data or material they possess as part of the competition gets leaked? And if the students were attacking a target inside a cyber range, how could this cause “loss or harm to the organizer and the country,” per the wording in the pledge?

The competition was held on a weekend over the New Year’s holiday last year, on December 30 and 31. If the target was indeed a real-world network, the researchers note that holidays are often a time when security teams are short-staffed or less attentive and alert to intrusions. About 200 students from 29 schools participated in the weekend competition, which consisted of three parts, according to a description of the contest published by the school: a theoretical knowledge competition, a vulnerability discovery contest, and a “public-network target, actual combat attack competition.” The latter is the portion that Cary and Benincasa believe focused on a real-world target.

The researchers surveyed more than 120 capture-the-flag competitions organized in China since 2004—many of them held annually—to understand how the country has used such competitions to recruit talent and expand its cyber offensive and defensive capabilities. They say that China really began to focus on developing its cyber talent in 2015, after the Edward Snowden leaks had exposed the extensive hacking operations conducted by the US National Security Agency for intelligence purposes.

To strengthen its cyber defenses, China focused on revamping and expanding cybersecurity education programs and recruitment efforts between 2015 and 2021. A big part of the latter included the development and regulation of capture-the-flag competitions.

CTFs aren’t unique to China, of course; they are held around the world, including in the US, where one of the oldest—held annually at the Defcon hacking conference in Las Vegas—has existed for nearly two decades. The US government also hosts hackathons to help recruit talent, though not at the scale of China.

Since 2014, the researchers say, more than 540 capture-the-flag rounds of competition have occurred in China. The most intense activity began in 2018, which is also when China’s Central Cyberspace Affairs Commission and Ministry of Public Security issued a notice about the regulation and promotion of such contests. Now Chinese nationals must apply to the MPS for permission to participate in hacking competitions outside of China. Chinese leaders noticed Chinese students and security professionals were having great success competing in hacking competitions internationally and realized that each time someone from China won one of these competitions, the country was losing a potentially valuable asset with the vulnerabilities they disclosed as part of the contest. Now any vulnerabilities discovered by a Chinese national must be reported to the government and not disclosed publicly.

But importantly, China’s efforts in building its domestic capture-the-flag contests means that today it has one of the most robust hacking contest ecosystems in the world, the researchers say, including sector-specific ones for health care and law enforcement. The researchers estimate that about 300,000 people have participated in the contests over the last decade, but the researchers found only four that are open to students. The contests are led by universities across the country while also being supported by companies and the government. Contest winners and others who demonstrate good skill get entered into a national database.

Cary says that if the Zhujian Cup did involve a live target, it’s “remarkable” that the organizers put students in a position that could make them legally culpable if they got caught. But if so, it wouldn’t be the first time the government used students for intelligence operations. In 2022, the Financial Times reported that China had recruited unwitting university students to translate documents and data stolen in espionage operations conducted by the country’s hacking teams, reportedly without telling the students the source of the material.

This also may not be the first time a hacking competition in China played a role in a real-world breach subsequently attributed to China. In 2015, researchers at Threat Connect found curious evidence of a possible overlap between the TOPSEC Cup hacking competition coordinated by China’s Southeast University in 2014 and the subsequent hack of health care giant Anthem that same year. Southeast has financial connections to China’s civilian intelligence agency, the Ministry of State Security.