Lessons from the iSOON Leaks



  • A Chinese Ministry of Public Security (MPS) contractor called iSOON (also known as Anxun Information) that specializes in network penetration research and related services has had its data leaked to GitHub.
  • Based on the level of detail, leaked chat logs, amount of data, and corroboration from overlaps indicators of compromise (IOCs), there is a high level of confidence it is legit.
  • Preliminary findings from less than one week since the leak revealed that it contains unprecedented insights into how the Chinese MPS operates by using Chinese commercial surveillance vendors and what their technical capabilities are.
  • The Chinese MPS is China’s internal security service that primarily focuses on internal and border security, counter-terrorism, surveillance. The MPS is comparable to the Russian FSB, the US DHS or the UK’s MI5.
  • The most interesting findings have come from iSOON’s product whitepapers and confidential slide deck presentations given to their MPS clients.

About the iSOON Leak

The iSOON leak has revolutionized the open source intelligence (OSINT) community's general understanding of the Chinese MPS' cyber operations. Its significance is comparable to the NTC Vulkan leak that impacted the Russian military intelligence (GRU) and Russian federal security service (FSB) as well as the Snowden Leak or the Shadow Brokers Leak, both of which impacted the US National Security Agency (NSA).

The leak is around 170MB and contains images of documents in Mandarin. They were machine translated to English for this blog. Therefore, not all of the facts will be available for a some time as the community digs in and there may be some misunderstandings in this blog due to not being a Chinese-speaking analyst myself.

The company, iSOON has had a long development history. Established in 2010, iSOON has been involved in various activities related to "network penetration research" and "overseas special case network work" and has conducted national training programs on network security. They have received commendations from CCP ministerial and central leaders for their contributions. In 2019, iSOON was selected as one of the first units installed by the Cyber Security Bureau of the MPS.

On 16 February 2024, a GitHub repository ("github[.]com/I-S00N/I-S00N") containing the data dumped from iSOON was uploaded to the site by an unknown source: 

And also on Twitter:

iSOON's Services

According to an iSOON presentation in the leak, the structure of firm includes three penetration teams, one security research team, and one basic support team, comprising a total of around 70 people.

Their primary targets are countries and regions such as Central Asia, Southeast Asia, Hong Kong, Macao, and Taiwan. iSOON offers services such as "APT Service System," "Target Penetration Services," and "Battle Support Services" to their MPS clients. 

They have also reportedly targeted various regions in China, including Xinjiang and Tibet, which includes dissidents, illegal gambling rings, illegal pornography rings, as well as illegal pyramid schemes (which all come under MPS jurisdiction).

Notably, iSOON featured in a slide deck their attacks on the Indian and Nepalese governments, targeting government departments such as the Ministry of Foreign Affairs, Defense, Home Affairs, Finance, and the Nepalese Presidential Palace.

iSOON's Victims

Fellow security researcher @azakasekai_ also was first to find that the leak contains conversations and spreadsheets about iSOON’s victims, including how many terabytes of data has been exfiltrated from certain government entities, telecommunications firms, medical organizations, and academic sectors of countries such as: Pakistan, Kazakhstan, Kyrgyzstan, Malaysia, Mongolia, Nepal, Turkey, India, Egypt, France, Cambodia, Rwanda, Nigeria, Hong Kong, Indonesia, Vietnam, Myanmar, Philippines, and Afghanistan.

Examples of some of the victim organizations include:

  • Ministry of Foreign Affairs, Myanmar 
  • National Taiwan University Hospital
  • Tamkang University, Taiwan
  • National Intelligence Agency, Thailand
  • Ministry of Foreign Affairs, Thailand
  • Ministry of Health, Rwanda 
  • Apollo Hospital, India
  • Punjab Anti-Terrorism Center, Pakistan
  • Beeline Communications, Kazakhstan
  • Tele2 Communications, Kazakhstan
  • Skytel Communications, Mongolia
  • Ministry of Public Security, Mongolia
  • Nepal Telecom
  • Chinese University of Hong Kong
  • Hong Kong Shue Yan University
  • Tung Wah College, Hong Kong
  • Paris Institute of Political Studies, France

Fellow security researcher @haxrob also noticed that some of the leaked files in the GitHub repository also contain call detail records (CDR) and location based services (LBS) related systems stolen from telecommunications entities. This highlights that the iSOON operators compromised telecommunications entities with the aim to obtain subscriber metadata to support intelligence collection objectives, in support of Chinese MPS requirements. With persistent access to LBS records it would be possible for iSOON and MPS operators to perform real time lookups of a subscriber's location, quite possibly to a very high degree of accuracy.

iSOON's Products

A range of products and tools are offered by iSOON. This includes a "Twitter Public Opinion Guidance and Control System," for monitoring dissidents; custom remote access software for Windows, Linux, macOS, iOS, and Android; custom hardware for "WiFi Proximity Attack Systems," disguised as Xiaomi battery packs; an "Email Analysis Intelligence Decision-Making Platform," an "Individual Tool Box," consisting of various hacking tools; a "Microsoft Email Encryption" platform, for email exfiltration and analysis; and an "Automated Penetration Testing Platform," that combines various public hacking tools into a Software-as-a-Service (SaaS) platform.

Twitter Public Opinion Guidance and Control System 

The Twitter Tool whitepaper in the leak was used by iSOON to sell its commercial surveillance platform to the Chinese MPS for monitoring dissents. Notably, the iSOON developers also claimed to have a 1-click exploit to bypass Twitter two-factor authentication (2FA) security controls to gain control over the target’s account. This exploit was to be distributed via Twitter direct messages (DMs) in the form of URLs, which iSOON called forensic links. These forensic links can gain access to the accounts but also gather IP addresses, IP locations, device type, and browser version.

The Twitter Tool whitepaper also showed in detail how the Chinese intelligence services access Western social media services and bypassing the Great Firewall of China. The platform enabled the Chinese MPS to view tweets and Twitter account profiles of target dissidents. The Chinese MPS operators using the iSOON platform could obtain "evidence" using the collection system in the platform, which involves indexing content from Twitter users criticizing the Chinese government.

Email Analysis Intelligence Decision-Making Platform

The Email Analysis Intelligence Platform developed by iSOON was sold to the Chinese MPS to support intelligence production from massive amounts of stolen emails. The platform was designed for mass email data analysis that enabled them to mine mailboxes and conversations for personal information, as well as IP addresses from email headers and extract details from email attachments. It boasted that it can handle terabytes of email data and index it for performing keyword searches and performs automatic email translation from any language.

The Email Tool whitepaper included the architecture of the system that enables mass email data analysis using machine learning (ML) algorithms, optical character recognition (OCR), and other methodologies.

One of the notable screenshots from the Email Tool whitepaper showed that the platform can also enable MPS operators to recreate their target’s inbox and read through their stolen emails 

Automated Penetration Testing Platform

Another platform developed by iSOON for the Chinese MPS is used for offensive cyber operations against a range of targets, like those mentions in the iSOON services section above. The automated penetration testing platform is made of various modules built using custom and publicly available tools. It offers capabilities for all sorts of offensive techniques (shown below), such as running phishing campaigns, application exploitation, cross-platform payload generation, and RATs, which includes legitimate remote monitoring and management (RMM) tools.

iSOON Overlaps

Several overlaps and connections associated with iSOON and known threat groups have already been discovered. The company already had a few ties to known Chinese advanced persistence threat (APT) groups such as Chengdu 404 (linked to APT41). Additional ties were uncovered through pivoting on indicators of compromise (IOCs) present in the leak, which adds further legitimacy to the validity of the leak.

The connection to threat group called POISON CARP was via an IP address (74.120.172[.]10) that appeared in the iSOON data leak that hosted a phishing site (mailnotes[.]online). The phishing site was observed as one of the IOCs in CitizenLab's “Tibetan Groups Targeted with 1-Click Mobile Exploits” report, which is in-line with the Chinese MPS’ operations supported by iSOON. 

The connection identified between iSOON and another Chinese commercial spying firm called Chengdu 404 was identified in Chinese court documents after the sanctioned firm Chengdu 404 sued iSOON over an intellectual property dispute.

The connection between iSOON and an APT group CrowdStrike track as JACKPOT PANDA was identified through another IP address (8.218.67[.]52) that appeared in the data leak. This IP address was referenced in Trend Micro's report “Probing Weaponized Chat Applications Abused in Supply-Chain Attacks” which appeared to be focused on targeting the online gambling industry, something iSOON also referenced as a target sector in their services slide deck.

Further pivots on some of the IP addresses spotted in the screenshots of the custom Windows malware in the iSOON product whitepapers as well as an overlapping reference to “TreadStone” in US Justice Department’s indictment of APT41 and Chengdu404. This revealed ties to the infamous ShadowPad and Winnti malware families. These custom Windows malware families have been tied to numerous Chinese cyber-espionage campaigns and used by multiple Chinese APT groups.

Using ACH to judge who was responsible for the iSOON leak
Note: This section of the blog is aimed towards my readers who are CTI professionals as well as those interested in the process of making intelligence assessments. Feel free to skip this part.

The answer, at the time of writing, is... we simply don't know. However, this is a good opportunity to perform what we call in the intelligence biz Analysis of Competing Hypothesis (ACH) to mitigate our unconscious biases when making an assessment. ACH was developed by Richards Heuer a CIA veteran analyst, who wrote the Psychology of Intelligence Analysis. It can help make judgments that entail a high risk of error in reasoning (in my case, I have only a few years experience investigating Chinese APT groups). The goal of ACH is to help an analyst overcome/minimize, some of my cognitive limitations to provide prescient intelligence analysis.

The three main steps to ACH is Hypothesis Generation (identify all possible explanations), Evidence Collection (list arguments for and against), and Diagnostics (test the arguments against each hypothesis). ACH encourages an analyst to consider one piece of evidence at a time, and examine it against all possible hypotheses. We then refine and prioritise our hypotheses before making a final conclusion. To perform an ACH, a Matrix is needed with evidence on the horizontal and our hypotheses on the vertical. This ACH Matrix will be used to determine which data points are the most helpful in assessing the likelihood of the presented hypothesis.

The first step involves developing our hypotheses based on the currently available intelligence. Generating as many hypothesis as necessary is best practice to ensure the inclusion of all available evidence. Ideally, when new evidence becomes available, all of our hypotheses should be evaluated again. 

List of hypotheses of potential candidates for the iSOON leak:
  1. An iSOON Revengeful Ex-Employee
  2. A rival Chinese Contractor
  3. A rival Chinese Agency to the MPS
  4. A Foreign Intelligence Agency
  5. An Anti-CCP Hacktivists
  6. Chinese Cybercriminals
In an ACH Matrix, a "+" sign means the evidence may support the hypothesis and a "-" sign indicates it may not. A "0" sign means it neither supports nor disproves the hypothesis. We then use the ACH Matrix to prioritise the hypothesis with the most supporting evidence

We then need to determine evidentiary dependence. This involves checking if our leading hypothesis has been prioritized based on critical evidence that, if deemed inauthentic and dismissed later on, would cause us to reconsider our entire assessment.

In the iSOON leak case, one of the most telling pieces of evidence that it came from a revengeful ex-employee was the headings of the files in the GitHub repo display a common sentiment that iSOON was a bad company to work for. The other key pieces of evidence that support it was a revengeful ex-employee include that they had the permissions and access to the resources and showed that their employees were working in less than ideal conditions, which involve difficult-to-perform tasks, such as domestic surveillance that supports arrests and ethnic cleansing of their fellow citizens.

Another valid hypothesis worth considering were whether a foreign intelligence agency was involved in the leak. However, key evidence that goes against this hypothesis is that the intelligence revealed in this leak would be much more valuable for a foreign agency to exploit rather than leak to the world and other intelligence agencies, who are undoubtedly studying this leak as well (and potentially this blog too! Hi guys 👋). The other arguments against the idea that a foreign agency was responsible is the potential danger it could cause, such as putting foreign nationals in China in physical danger or create escalations with China who would likely retaliate with a targeted hack-and-leak operation of their own. There is ultimately little benefit for a foreign agency to leak this information if they had managed to steal it themselves from iSOON's network.

The final hypothesis that seasoned analysts would be interested in discussing is whether an anti-CCP hacktivist group like IntrusionTruth was involved in this leak. However, in a very un-hacktivist fashion, nobody has publicly claimed responsibility for this leak, as a victory. Hacktivism is often described as a cyber-enabled influence operation and by leaking it anonymously, it goes against this common behavioural trait of theirs.


The iSOON leak is one of the most significant cyber threat updates related to China in recent years. The tools offered by iSOON and campaigns run by their operators highlight how both the Chinese MPS and Chinese Ministry of State Security (MSS) outsource their intelligence gathering to commercial surveillance vendors.

The links already uncovered between multiple long-running APT campaign and iSOON as a single entity has essentially taken a hammer and smashed the notion of neatly defined “threat groups” conducting campaigns in a siloed manner. The leak reinforces the idea that APT groups in China are connected to each other in many ways like the cybercrime underground or even the Western cyber defense industry in many ways than we admit.

Overall, I anticipate a lot more interesting findings and analysis to arise from the iSOON leak. I recommend all CTI teams to keep an eye out for mentions of iSOON as more revelations are bound to appear. 

Further Reading:

Popular posts from this blog

Raspberry Robin: A global USB malware campaign providing access to ransomware operators

Tracking Adversaries: Scattered Spider, the BlackCat affiliate