A recent report by CrowdStrike Intelligence has shed light on a significant gap in enterprise security, with the identity and access management (IAM) pivot attack chain posing a $2 billion threat to cloud security. This attack chain involves recruitment fraud, where developers are tricked into installing malicious packages that exfiltrate cloud credentials, allowing adversaries to gain access to cloud environments within minutes.
The attack chain is becoming increasingly sophisticated, with threat actors using social platforms like LinkedIn and WhatsApp to deliver trojanized packages. In one case, a European FinTech company was targeted through recruitment-themed lures, resulting in the diversion of cryptocurrency to adversary-controlled wallets. The Cybersecurity and Infrastructure Security Agency (CISA) and security company JFrog have tracked overlapping campaigns across the npm ecosystem, with JFrog identifying 796 compromised packages. The research highlights the need for runtime behavioral monitoring, as dependency scanning alone is not enough to detect credential exfiltration.
According to Google Cloud‘s Threat Horizons Report, weak or absent credentials accounted for 47.1% of cloud incidents in the first half of 2025, with misconfigurations adding another 29.4%. The report emphasizes the importance of monitoring identity behavior inside cloud environments, rather than just authenticating successfully. CrowdStrike‘s SVP of intelligence, Adam Meyers, noted that decentralized currency is ideal for attackers, as it allows them to avoid sanctions and detection simultaneously. The company’s field CTO of the Americas, Cristian Rodriguez, explained that revenue success has driven organizational specialization, with threat groups splitting into distinct units targeting cryptocurrency, fintech, and espionage objectives.
The attack chain has significant implications for AI infrastructure, with OpenClaw, an open-source autonomous AI agent, connecting to email, messaging platforms, calendars, and code execution environments through model context protocol (MCP) and direct integrations. Cisco‘s AI security research team has warned that this tool is a “groundbreaking” capability from a security standpoint, but also an “absolute nightmare”. The IAM implications are direct, with a successful prompt injection against an AI agent potentially leading to automated lateral movement and access to APIs, databases, and business systems.
To mitigate these risks, enterprises need to deploy runtime behavioral monitoring, ITDR that monitors identity behavior across cloud environments, and AI-specific access controls that correlate model access requests with identity behavioral profiles. As PwC‘s deputy leader for cyber, data, and tech risk, Morgan Adamski, noted, getting identity right, including AI agents, means controlling who can do what at machine speed. The Nvidia and Ring are not mentioned in the source text, however OpenAI is not mentioned as well, but the article discusses AI in general.
The report concludes that the perimeter is no longer the primary battleground, with identity being the new frontier in cybersecurity. Enterprises must audit their IAM monitoring stack against the three-stage attack chain, ensuring they have the necessary controls in place to detect and prevent these types of attacks. As Sectigo‘s senior fellow, Jason Soroko, noted, the root cause of these issues is often a refusal to master security fundamentals, such as validating credentials and implementing proper access controls.
Leave a Reply