Team Cymru found 179 malware-pattern servers, and no named operators
Team Cymru's Nov. 17 financial-sector espionage research matched 64 ShadowPad and 115 Vshell servers between January 1 and June 1, 2026, while declining to attribute them to named threat groups.
By Olivia Hart
4 min read

What's News
- Team Cymru identified 64 servers matching known ShadowPad command-and-control patterns between January 1 and June 1, 2026.
- The same window produced 115 servers matching Vshell patterns, 179 matches in total.
- The research, published November 17, examines China-nexus espionage aimed at financial services.
- Team Cymru says it has not attributed the matched servers to any named threat group.
- CrowdStrike's 2026 Financial Services Threat Landscape Report names HOLLOW PANDA, GENESIS PANDA and MURKY PANDA among China-nexus actors.
By Olivia Hart
Team Cymru identified 64 servers matching known ShadowPad command-and-control patterns and 115 matching Vshell patterns between January 1 and June 1, 2026. The company has not attributed those servers to any named group, and the counts measure infrastructure rather than compromised banks.
The research, published November 17, examines China-nexus espionage targeting financial services through the infrastructure behind malware and covert relay networks. It joins a sector-wide threat assessment with Team Cymru's independent tracking of internet infrastructure.
What did the servers actually count?
Combined, the two sets total 179 matches over a five-month window, roughly 36 a month. The Vshell count ran about 1.8 times the ShadowPad count. Neither figure is a tally of intrusions.
Team Cymru frames that limit as its central finding: "Detecting a known malware tool supports investigation without establishing which threat actor operates it."
Why financial services?
CrowdStrike's 2026 Financial Services Threat Landscape Report identifies China-nexus actors as a continuing intelligence threat, focused on personally identifiable information and economic data. Financial organizations therefore face a wider problem than attacks built to steal money or demand a ransom.
The target may be what a financial institution knows: its customers, commercial relationships and sensitive transactions. For defenders, immediate financial loss is an incomplete measure of an incident's significance. A quiet account compromise can warrant attention even while payments continue normally.
Can the servers be tied to HOLLOW PANDA, GENESIS PANDA or MURKY PANDA?
Team Cymru cites CrowdStrike's reporting on HOLLOW PANDA using ShadowPad, GENESIS PANDA using Vshell and Fscan, and MURKY PANDA using operational relay box networks to reach Microsoft 365 accounts. Its own infrastructure tracking does not independently assign those observations to the named groups.
MITRE ATT&CK describes ShadowPad as a modular Windows backdoor first identified in the 2017 NetSarang supply-chain compromise. Initially associated with APT41, it has since appeared in activity involving several Chinese threat groups. Documented capabilities include downloading additional code, collecting system information and communicating through several protocols.
A malware family, an infrastructure cluster and a threat actor are three different investigative objects. Connecting them requires corroborating evidence, and Team Cymru says so plainly: "Shared malware families and infrastructure clusters require corroborating evidence before analysts can attribute them to actors."
How do relay networks blunt blocklists?
Operational relay box, or ORB, networks combine rented virtual servers with compromised routers and other internet-connected devices, passing traffic through multiple nodes before it reaches the target. The target may see the final relay rather than the operator's origin.
Rotating exit nodes make static blocklists less durable, and the use of residential or business broadband risks blocking legitimate users alongside suspicious activity. The mesh also creates investigative openings: relationships between nodes, unexpected communications and distinctive infrastructure characteristics. The useful unit of analysis becomes connected activity rather than one IP address in isolation.
Team Cymru's warning applies to defenders running blocklists today: "Relay networks blending rented servers with compromised devices make static blocklists less durable for defenders."
What does network-flow intelligence add?
Network-flow records show which systems communicated and when. The IETF's IPFIX standard defines how measured flow information moves from exporters to collectors, allowing tools to organize observations about network activity. Analysts correlate such records with other telemetry; they do not supply an automatic explanation of purpose.
Team Cymru's platform combines that network-flow visibility with passive DNS, routing information, certificates and malware intelligence. Its NetFlow offering emphasizes external reconnaissance, command-and-control tracking and historical investigation. The outside vantage point complements an institution's own logs: internal telemetry shows what happened to an account or endpoint, while external observations supply context about the infrastructure it contacted. Coverage and corroboration still determine how far an analyst can take the conclusion.
Where does AI fit, and what isn't shown?
Team Cymru offers an MCP server connecting AI agents to Pure Signal intelligence. Model Context Protocol provides an interface through which an assistant can request external information, bringing infrastructure enrichment into an AI-assisted workflow and cutting repetitive lookups during an investigation.
The value depends on keeping source evidence attached to the analysis and preserving the distinction between an observed connection and an attribution judgment. The financial-services research does not establish that the reported intrusions themselves used AI. That is a product capability, not a finding about the intrusions.
What should defenders do with this?
Joint guidance from CISA and partner agencies recommends strong network-flow monitoring, investigation of unauthorized device-configuration changes and patching vulnerable equipment. Although written for communications infrastructure, the guidance also addresses organizations operating enterprise network equipment.
For financial institutions, that supports an investigation connecting perimeter activity, network changes and account events. A suspicious relay becomes more actionable when placed on a timeline with a VPN session, an unexpected configuration change or unusual access to a sensitive account.
Team Cymru's findings argue for infrastructure tracking alongside malware detection and identity monitoring, with conclusions held to what the investigation actually establishes. Until analysts attach corroborating evidence to the 179 matched servers, they remain infrastructure observations rather than identified operations.
Source: unite.ai
More from Olivia Hart
Show full bio
Staff writer covering people, launches and day-to-day practice in entrepreneurship, startups and venture funding at Business Bearings.
2,378 articles