AuditTeam disguises ransomware extortion as a paid security audit
AuditTeam has used its OpsAgent malware framework to compromise organizations in at least seven countries and territories, pairing data theft and file encryption with an unusual extortion pitch: pay an “Audit & Consulting fee” and receive a vulnerability report instead of having stolen data published. The campaign was active from January through at least October 2026, according to the reported investigation.
Dozens of organizations were affected in Russia, Italy, the Philippines, Hong Kong, Senegal, Finland and China. AuditTeam’s leak site also listed alleged victims in seven other countries. The group’s branding does not make the activity a legitimate audit: the ransom note is associated with a cryptor module, meaning files may already be encrypted when victims encounter the demand.
A modular implant with expanding capabilities
OpsAgent is a multi-function implant that supports host reconnaissance, file and process operations, privilege escalation, keylogging, screenshots, shellcode injection and network pivoting through linked agents. Its operators can deliver additional payloads remotely, using named pipes to pass encrypted shellcode to background jobs. The framework’s commands also support file transfer, proxy connections and forwarding tasks between compromised systems.
Researchers observed multiple loader types during the campaign. Earlier C and .NET loaders inject the unpacked payload into another process, while a Rust loader discovered in September 2026 retrieves an encrypted payload and executes it within its own process. The implant itself was identified as version 5.2.8. A Rust rewrite retained much of the earlier functionality and added capabilities including interactive shells and changes to the command-and-control (C2) communication process.
In the Rust version, initial C2 communication uses HTTP POST requests. An upgrade command can move the implant to a persistent WebSocket connection over TLS. Earlier versions also communicate over WebSocket and can obtain C2 addresses from a public GitHub repository. The repository contained encrypted configuration data, and the Rust version changed the configuration encryption scheme from AES-CBC to AES-GCM. The repository was removed in November 2026, but investigators found recent Rust implants configured not to use the GitHub lookup feature.
Defense evasion and encryption
Two specialized modules increase the impact of an intrusion. An AV/EDR-killer module installs a digitally signed driver presented as a Beijing Jiangmin Antivirus component. The driver can terminate processes by filename, with a target list of about 220 entries that includes security products, monitoring agents, remote-access tools, backup services and database software. A valid signature can help a driver run, but it does not establish that its use is authorized or safe.
The cryptor supports AES-CBC or ChaCha20 encryption, with X25519 key exchange used to protect file keys. Before encryption, it can terminate database processes and stop backup services such as Veeam. It can also remove shadow copies, empty the recycle bin and clear temporary folders. The ransom note threatens public disclosure of stolen data if the fee is not paid, while offering the purported vulnerability report as the alternative.
AuditTeam’s approach combines ransomware impact with a claim of security remediation. That framing may complicate initial incident handling if staff treat the demand as a conventional consulting offer. Organizations should instead respond as to a suspected intrusion and extortion incident, preserving evidence and involving incident-response and legal teams.
Detection and response considerations
The reported infrastructure includes AWS CloudFront subdomains used as C2 endpoints. Because CloudFront addresses are shared infrastructure, the investigation cautions that underlying IP addresses are not reliable standalone indicators for blocking. AuditTeam also rotated domains and used a public repository to publish C2 configuration, so defenders should not rely on one infrastructure indicator as the sole detection method.
Security teams can use the reported file hashes, paths and domains as investigative leads, while validating them against current telemetry and threat intelligence. The listed paths include several locations for a file named syscache.dat, as well as kvcore.sys and kv_dbg.log. Endpoint teams should review suspicious driver installation and process-termination activity, especially where security, backup or database processes are targeted. Network monitoring for unusual agent communications, remote payload delivery and unexpected WebSocket or HTTP activity may provide additional context.
The campaign demonstrates how a custom implant can support several stages of an intrusion, from access and lateral movement to defense disruption and encryption. AuditTeam’s “audit” language should not obscure the underlying risks: data theft, operational disruption, loss of recovery options and public disclosure.