AuditTeam OpsAgent ransomware presents extortion as a compulsory security audit. Kaspersky’s November 18, 2026 research links the campaign to AuditTeam with high confidence and describes infections from January through at least October 2026. Its analysis separates a general-purpose implant from two modules that disable defenses and encrypt files. Responders should preserve endpoint, service and backup evidence before relying on any single ransom note or network indicator.
The group’s note demands an “Audit & Consulting fee” and promises a vulnerability report if the victim pays. Kaspersky says the note is placed by the encryptor, so its appearance means files have already been encrypted. The group also threatens public disclosure of stolen data. That threat is the actor’s claim; the report does not give a verified count of successful data disclosures.
How the AuditTeam OpsAgent Ransomware Chain Unfolds
Kaspersky first found suspicious OpsAgent activity at an organization in Hong Kong on March 10, 2026. It later analyzed infections across seven countries and territories. The research describes the implant, loaders and destructive modules, but does not identify a common initial access method. Treat the following sequence as the observed toolchain, not a proven path into every affected network.
The loaders place OpsAgent into execution. Kaspersky found C and .NET variants that inject into another process, followed by a Rust loader that runs the decrypted implant within its own process. This difference matters during a hunt: a rule focused only on remote process injection could miss the later loader. The source dates those discoveries to April, June and September 2026, respectively.
Once running, the C++ implant accepts operator tasks for host reconnaissance, file management, process execution, credential-related activity, screenshots and keylogging. It can also link compromised hosts for movement inside a network. Kaspersky describes a worker thread that handles most commands and separate jobs for some tasks. These are capabilities of the framework; the report does not show that every capability was used at every organization.
OpsAgent can discover command infrastructure through a public repository and communicates over encrypted channels. Kaspersky observed CloudFront subdomains used as command endpoints, with infrastructure changes during the campaign. Because that CDN is shared, its underlying IP space is not a reliable standalone block target.
Defenders need correlated endpoint and network behavior instead of treating a shared address as proof of compromise. Compare those signals with the actor context in your ransomware intelligence workflow.
Kaspersky describes two specialized modules alongside OpsAgent. One module loads a digitally signed antivirus driver and uses it to terminate security processes. The other terminates selected database and backup services, removes shadow copies and encrypts files. Kaspersky’s analysis establishes these module functions, but it does not publish a universal ordering for every intrusion.
- LoaderC and .NET variants inject; the later Rust loader runs in-process
- Main implantOperator tasks include reconnaissance, file access and movement
- Specialized modulesA signed driver can terminate security processes; a separate encryptor can stop services and encrypt files
- Evidence checkCompare driver loads, service stops and encrypted files; the report gives no universal module order
Why the C++ to Rust Change Alters Detection
The Rust rewrite preserves much of the original framework while changing execution and communications. Kaspersky says it adds interactive shells, directory creation and sleep interval changes. It can begin with HTTP requests and switch to a persistent WebSocket connection after an upgrade command. Its command identifiers are numeric rather than the C++ version’s text names.
The newer version also changes how it may retrieve command infrastructure. A configuration flag controls whether it checks a public repository for updated addresses, and its encrypted configuration uses a different scheme. Kaspersky reports that the repository was removed in November 2026. The latest Rust samples it found had repository checking disabled, so that removal does not establish the end of the campaign.
For defenders, these changes argue for detection across behavior rather than one loader or command pattern. Review unusual process creation and module loading, persistent outbound connections and service disruption as related evidence. The Rust version also removes the older pivot command set, according to Kaspersky. A hunt that expects those exact commands could therefore undercount newer activity.
Where the Disruption Becomes Visible
Kaspersky says the defense-disabling module installs a validly signed driver associated with Beijing Jiangmin Antivirus. Its process target list contains about 220 entries and spans security tools, monitoring agents, remote-access tools, backup services and database processes. A valid signature alone is therefore a poor reason to trust driver activity in this context. Review unexpected driver installation alongside the services and processes that stopped soon afterward.
The separate encryptor supports AES-CBC or ChaCha20 and uses X25519 for key exchange, according to Kaspersky. Its operational effect matters more to a responder than the cipher choice. It can stop database processes and backup services, delete shadow copies, encrypt selected files and leave the audit-themed ransom note. It also overwrites content before some deletion, which raises the cost of recovering evidence from the affected host.
Build a timeline from service-control events, driver loading, endpoint protection alerts, backup job failures and the first encrypted files. Check whether database services stopped before encryption and whether shadow copies disappeared. These checks are our response priorities derived from the module behavior in Kaspersky’s report. They are not a claim that a particular reader’s environment was compromised.
The incident should enter the organization’s ransomware response playbook when those signals appear together. Keep immutable backup and management-plane logs outside the affected host where possible. If security agents stop unexpectedly, collect independent telemetry before reinstalling them. A restored agent may hide the sequence of events if its prior service state is lost.
What the Reach Figures Actually Establish
Kaspersky reports dozens of infected organizations in Russia, Italy, the Philippines, Hong Kong, Senegal, Finland and China. Its observed set covers multiple industries, without a sector breakdown. The group’s leak site also lists seven additional countries, but those listings are actor claims. They must not be counted as confirmed infections or added to the researcher’s observed total.
From August 20 through November 18, 2026, a Brandefense Blackhole search for “AuditTeam” counted 166 dark-chat posts. The unit is posts matching that term, not organizations, infections or unique incidents. Some discussion may repeat the same underlying event. This volume supports monitoring the actor’s conversation, but it cannot validate a leak-site claim or date a new compromise.
Kaspersky’s high-confidence AuditTeam attribution rests on the audit-fee language, the group’s leak site and its described operating model. That assessment belongs to Kaspersky. A defender can use the pattern to prioritize collection, while keeping a local attribution finding separate from observable host evidence. Do not infer successful exfiltration solely from the note’s disclosure threat.
How to Prioritize the Response
Start by identifying hosts where the security stack, monitoring or backup services stopped unexpectedly. Correlate the stop times with new driver loads and command activity. Then check database services, shadow-copy events and file changes on those hosts. This order follows the modules’ potential to blind defenders before encryption.
Preserve logs from endpoint controls, backup consoles, database hosts and network gateways. A CloudFront address by itself is weak evidence because legitimate traffic shares the infrastructure. Pair network anomalies with process, service and driver events. The EDR-killer hunt plan offers a related way to organize that evidence, though it describes a different operation.
Test recovery from copies that the affected hosts could not alter. Verify whether backups completed before the service interruptions, and document any gaps. If encryption is confirmed, follow the isolation and restoration sequence in the response plan. The ransomware readiness page explains how to prepare those roles and recovery paths before a crisis.
For threat intelligence teams, maintain a separate list of observed infections, actor claims and local signals. Such monitoring can support the response, while local incident decisions must rest on evidence from the affected environment. The research does not establish a shared initial access route, a sector distribution or a verified total of data disclosures. Those gaps should remain open questions in the case record.
Frequently Asked Questions
What does the forced audit ransom note mean?
Kaspersky says the encryptor leaves the note after encrypting files. The fee and promised vulnerability report are the group’s extortion framing.
Does the removed public repository mean the campaign ended?
No. Kaspersky found newer Rust implants with the repository-check feature disabled, so its removal does not establish an end date.
Which evidence should responders review first?
Correlate unexpected driver loads, stopped security and backup services, database disruption, shadow-copy deletion and file encryption. This sequence is our defensive reading of the modules Kaspersky analyzed.
Are all countries on the leak site confirmed victims?
No. Kaspersky observed infections in seven countries and territories. The seven additional countries listed by the group are its claims, not independently confirmed cases.
