"This is the first time that we’re seeing an agentic adversary nation-state threat actor at this scale, and I think that this is probably, unfortunately, something we’ll see more of," Adam Meyers, senior vice president of counter adversary operations at CrowdStrike, told CyberScoop.
Who CrowdStrike says is behind the campaign: Vapor Panda and ARTEX
CrowdStrike attributes a rapid, broad-access intrusion campaign to a China state-backed threat group it tracks as Vapor Panda. According to CrowdStrike’s report shared with CyberScoop, Vapor Panda has been using ARTEX — described in the report as an automated agentic AI penetration testing framework — paired with a tool called DeepSeek to automate and accelerate exploitation across large numbers of targets.
Speed and scale: 140 sensitive targets compromised in 11 days
CrowdStrike says the campaign began on or after Oct. 17 and, within an 11-day window, compromised about 140 sensitive targets. The company reports the group targeted more than 200 organizations in at least 14 countries. CrowdStrike told CyberScoop that about 70% of Vapor Panda’s targets were confirmed compromised or had evidence indicating moderate confidence of a successful intrusion.
Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scrambleTargets named in the report
- Diplomatic and government targets: foreign ministries, a European country’s presidency and prime minister’s office, and defense and intelligence agencies.
- United States-based organizations: 19 U.S. entities in the pharmaceutical, semiconductor, defense, nuclear and financial sectors.
- Telecommunications: providers in the United States, Belgium and the Philippines.
CrowdStrike characterizes the campaign as a "diverse operational mission" focused on sectors and geographies that "Beijing considers economically or politically important," a description attributed to Meyers in the report.
How ARTEX and DeepSeek changed operational tempo
The report draws a sharp contrast between traditional "hands-on-keyboard" intrusion activity and agentic AI–driven operations. Meyers told CyberScoop that hands-on-keyboard activity tracked by CrowdStrike over the past 15 years typically involved up to 40 commands per hour. With ARTEX, CrowdStrike observed activity "exceed 1,500 commands an hour." The agentic tooling enabled concurrent attack paths and extremely frequent requests — reported as often as every 10 seconds — enabling the adversary to "hit more targets in a faster period of time."
CrowdStrike also notes the adversary’s ability to "drop a websell" (as stated in the report), and that deploying ARTEX across hundreds of targets makes a mass-exploitation campaign relatively inexpensive for an operator who can automate breadth and persistence.
Context: other ARTEX activity and the rising tide of agentic adversary attacks
CrowdStrike’s findings are positioned separately from ARTEX activity Microsoft Threat Intelligence attributed to Storm-0940 earlier in October, which included a supply-chain attack targeting ConnectWise customers. CrowdStrike emphasized that adversarial use of agentic AI is a rapidly growing trend: the firm observed roughly 45 agentic adversary attacks in September and reported 126 for October "and climbing."
As Meyers summarized, adversarial agentic AI use is "relatively new, yet also quickly becoming the new normal," and the increase in activity has prompted significant concern inside the security community. He warned, "We’re scrambling because we’re seeing so many of these agentic adversaries," and urged action, saying "It’s not hopeless, but you’ve got to take action. You can’t put your head in the sand."
What this means for technologists, policymakers, and affected enterprises
- Technologists and security teams: will need heightened visibility into endpoints, identities and cross‑domain telemetry — the exact capabilities CrowdStrike says are necessary to detect and respond to fast, agentic campaigns.
- Policymakers and regulators: face a growing pattern of automated nation-state activity that has already touched government offices and critical sectors; the report separates this campaign from other vendor-linked supply-chain incidents, underscoring multiple concurrent threats to consider.
- Affected enterprises and procurement leaders: organizations in pharmaceuticals, semiconductors, defense, nuclear, finance and telecoms should assume adversaries can rapidly scale exploitation attempts and prioritize detection of anomalous automated tasking and rapid command sequences across many assets.
The record CrowdStrike provided to CyberScoop shows a shift in malicious tradecraft rather than a new vulnerability class: existing intrusion techniques executed at machine speed and breadth through agentic orchestration. That combination — automated planning, sub‑second request cadence and concurrent attack paths across hundreds of targets — is the defining feature described in the report and the reason CrowdStrike and other defenders are treating the episode as a bellwether.
Read the original CyberScoop report: https://cyberscoop.com/china-vapor-panda-agentic-ai-artex-campaign-crowdstrike/
