UNC1069 Social Engineering of Axios Maintainer Led to npm Supply Chain Attack

Supply chain attacks have ceased to be the exclusive domain of automated scripts scanning repositories for vulnerabilities. The latest incident involving the extremely popular Axios library in the npm ecosystem proves that APT (Advanced Persistent Threat) groups are moving toward an operational model based on deep psychology and precise targeting of individuals. Behind the attack is the UNC1069 group, identified as an entity linked to North Korea, which has proven that the weakest link in the security of global IT infrastructure remains the human being, even if they are an experienced developer.

The scale of the problem is enormous, considering that Axios is one of the most frequently downloaded packages in the JavaScript world, used for handling HTTP requests. As the project maintainer himself, Jason Saayman, confirmed, the security breach was not the result of a code error or an API key leak, but the outcome of a weeks-long, carefully planned social engineering campaign. The attackers from UNC1069 demonstrated patience that goes beyond the standard framework of cybercrime, building a relationship based on trust before launching the final offensive.

Precise victim profile and mastery of manipulation

The attack on Jason Saayman was not accidental. The UNC1069 group conducted detailed open-source intelligence (OSINT) to tailor the conversation scenario directly to the developer's interests and responsibilities. According to Saayman's account, the attackers impersonated the founder of a fictional startup or technology project, offering a collaboration that at first glance appeared attractive and substantive. This is a "long con" approach, where hackers do not ask for access in the first message but build authority over days or even weeks.

Modern social engineering campaigns, such as the one conducted by UNC1069, exploit the mechanism of reciprocity and professional courtesy. Attackers often send code samples for review or ask for help in solving a purported technical problem, which is actually a Trojan horse. In the case of Axios, the goal was to obtain permissions allowing for the injection of malicious code directly into the official release of the package, which automatically spread the threat to millions of servers and client applications worldwide.

The role of AI in shortening defense response time

The analysis of the incident coincides with the publication of the Zscaler ThreatLabz 2026 VPN Risk Report, developed in collaboration with Cybersecurity Insiders. This document sheds new light on how technology is changing the dynamics of attacks. According to Zscaler experts, artificial intelligence (AI) has drastically shortened the so-called "human response window." Thanks to the automation of generating credible content and deepfakes, attackers can run dozens of such campaigns simultaneously while maintaining a high quality of manipulation.

The report highlights key threats resulting from the current model of remote work and infrastructure access:

• AI as a breach accelerator: Generative tools allow groups like UNC1069 to create linguistically and substantively perfect messages, making them harder to detect by spam filters and developer vigilance.
• Remote access as the fastest path to attack: Exploiting vulnerabilities in VPN systems or simply stealing credentials through social engineering remains the most effective method of infiltrating corporate networks.
• Erosion of trust in Open Source: The attack on Axios shows that trust in the maintainers of popular libraries is being systematically exploited by state-sponsored hacking groups.

Technical implications of the Axios package takeover

The takeover of publishing rights in the npm repository for a project as critical as Axios has catastrophic consequences. Every CI/CD system that automatically downloads the latest versions of dependencies becomes an infection vector. The attackers from UNC1069, after obtaining access from Jason Saayman, had the opportunity to introduce a payload that could steal environment variables, cloud access keys (AWS/Azure), or end-user data. This is a classic example of a horizontal attack, where one infected code distribution point hits thousands of organizations.

The nature of North Korean group activities often includes cryptocurrency theft or infiltration for industrial espionage. In the case of Axios, the precision of the attack suggests that the attackers were interested in long-term, undetectable access to data transmitted by infected applications. Using npm as a malware distribution platform is particularly dangerous because most security scanners focus on CVE-type vulnerabilities, rather than authentic—albeit malicious—code signed by the legitimate project maintainer.

The need to redefine the trust model in IT

The UNC1069 incident forces the technology industry to move away from a trust model based on the identity of a single human being. Even the most aware developer can fall victim to advanced manipulation, especially when it is supported by AI tools. A solution gaining importance is the introduction of mandatory multi-party authorization for the publication of critical packages and rigorous code signing in isolated environments.

"AI not only helps in creating code but drastically shortens the time in which a human is able to react to a threat, turning remote access into the fastest path to a security breach." — Zscaler ThreatLabz 2026 VPN Risk Report.

The evolution of UNC1069's methods shows that cyber warfare has moved to the level of human relationships. Companies must understand that their security depends not only on their own firewalls but on the psychological resilience of the open-source developers whose work they use. The lack of systemic solutions supporting maintainers of projects like Axios in the fight against social engineering will lead to further, even more severe breaches in the global software supply chain.

In an era of widespread use of AI to personalize attacks, traditional cybersecurity training is becoming insufficient. The industry must focus on Zero Trust architecture not only within corporate networks but primarily in software development processes. Only by technically limiting the ability of a single person to authorize changes in critical infrastructure can the risks posed by groups like UNC1069 be realistically minimized. The future of supply chain security lies in automatic code integrity verification that trusts no one—not even the lead developer.