Skip to main content

Hit by ransomware? Isolate affected systems now. Do not reboot or reformat.

SheMo Noransom舍末无勒

Ransomware family

INC Ransom Ransomware Decryption & Data Recovery

  • Active
  • Critical
  • No public decryptor

INC Ransom emerged in 2023 and has passed 800 claimed victims by 2026, making it a top-tier RaaS operation. Marked by the .INC extension and INC-README notes, it exploits Citrix and SonicWall edge flaws and ships an ESXi encryptor. No public decryptor exists.

First seen
2023-07
File extensions
.INC .inc
Ransom notes
INC-README.txt
Affected platforms
Windows / Linux / VMware ESXi

Family profile

File extensions
  • .INC
  • .inc
Ransom notes
  • INC-README.txt
  • INC-README.html
Contact patterns
  • Tor (.onion) negotiation portal + per-victim personal ID
  • INC Ransom leak site on Tor
  • Ransom notes mass-printed to victim printers and fax machines
Aliases / versions
INC Ransomware、GOLD IONIC、INC Ransom ESXi Locker、Lynx(源码衍生品牌)
First seen
2023-07
Status
Active
Threat level
Critical
Affected platforms
  • Windows
  • Linux
  • VMware ESXi
Tags
  • Leak-site regular
  • Active
  • Ransomware-as-a-Service
  • Double extortion
  • Targets virtualization
  • Exploits vulnerabilities
  • RDP brute force
Decryptor
No public decryptor

No free public decryptor exists for INC Ransom. The Windows build combines Salsa20/AES symmetric encryption with Curve25519 key agreement, while the Linux/ESXi build derives AES-128-CTR keys via X25519 ECDH, both with per-file keys and no published implementation flaw. Note also that INC's source code was offered for sale in 2024 and reused by crews such as Lynx, so different variants have independent key systems and any future break in one branch would not transfer. Recovery should centre on backups and snapshots, structural repair and unencrypted duplicates.

Latest activity

  1. SonicWall SMA 1000 zero-days CVE-2026-15409/15410, patched mid-July 2026, chain to RCE. INC became the dominant exploiter (~885 claimed victims), stealing credentials, session databases and TOTP MFA seeds.

    Sources
  2. The FortiBleed credential-theft operation hit 430,000+ FortiGate firewalls, with sniffers on ~19,000. SOCRadar found its browser sessions open INC and Lynx affiliate panels and negotiation chats.

    Sources
  3. Cyber Centaurs found INC had left hardcoded Restic backup credentials (keys, repo paths, S3 passwords) in victim environments, exposing reused exfiltration storage and recovering stolen data for 12 US organisations.

    Sources

Overview

INC Ransom first appeared in July-August 2023, running multi-layered extortion - encryption, leak and direct harassment - with healthcare, education, industrial manufacturing and government as early focus areas and victims overwhelmingly in Europe and North America. Its Linux variant was first observed in December 2023 with full ESXi support.

2024 was the pivotal year. Attack volume climbed sharply, several UK NHS organisations and a children's hospital were hit, drawing significant attention. At the same time, INC's source code - Windows and Linux/ESXi versions included - was offered for sale on underground forums for 300,000 dollars. The Lynx ransomware that surfaced shortly afterwards was shown by Unit 42 to share roughly 48% of its functions with INC overall and 70.8% among common functions, with even higher overlap in the Linux/ESXi variant, indicating Lynx either bought or repurposed the INC codebase.

By 2026 INC had grown into one of the most active operations in the ecosystem, with public trackers counting more than 930 claimed victims on its leak site and dozens still added each month, alongside observed large-scale exploitation of SonicWall SMA 1000 series flaws for initial access.

The relevance for Chinese organisations lies in the entry points. INC leans heavily on edge device vulnerabilities - Citrix NetScaler CVE-2023-3519 and CVE-2023-4966 ("Citrix Bleed"), SonicWall SMA series - and RDP logins with valid accounts. Manufacturers and hospitals in China run the same classes of remote access appliances with typically long patch cycles, so the exposed surface is identical. No public report confirms a targeted INC campaign against mainland China organisations.

How to identify it

Extension. Encrypted files get .INC appended (lowercase .inc in some builds) with the original name preserved, so ledger.xlsx becomes ledger.xlsx.INC. A characteristic signature block is appended to the end of each encrypted file and is a reliable attribution artefact.

Ransom notes. Both INC-README.txt and INC-README.html appear in affected directories. The English text carries a victim-specific personal ID and a Tor portal address, pressing with a claim that data has been downloaded and a 72-hour contact deadline.

Printed notes. One of INC's signature behaviours is pushing the ransom note to every reachable printer and fax machine, so staff arriving in the morning find it on paper as well as on screen - frequently how the incident is first discovered.

Other indicators.

  • Anomalous access and session hijacking against Citrix NetScaler or SonicWall SMA appliances beforehand.
  • NETSCAN.EXE sweeps, AnyDesk installations, and PsExec renamed (for example to winupd) for deployment.
  • SystemSettingsAdminFlows.exe abused to tamper with Windows Defender settings, and a custom av.exe terminating EDR processes.
  • MEGASync, Rclone, 7-Zip or WinRAR appearing on non-administrative hosts, indicating exfiltration.

ESXi. Bulk VM shutdowns via esxcli or vim-cmd followed by encryption of files under /vmfs/volumes.

Infection vectors

INC Ransom's access is concentrated on edge appliances and valid credentials - mature tradecraft applied repeatedly:

  • Edge device exploitation. Citrix NetScaler ADC/Gateway CVE-2023-3519 and CVE-2023-4966 ("Citrix Bleed", which hijacks sessions and bypasses MFA) are its signature entry points, joined in 2026 by large-scale exploitation of SonicWall SMA 1000 series flaws.
  • Valid accounts over RDP. Purchased or stolen credentials used for direct remote logins, skipping exploitation entirely.
  • Phishing. Loader delivery to establish an initial foothold.
  • Internal reconnaissance. NETSCAN.EXE, built-in commands and Active Directory queries to map the network and assets.
  • Lateral movement. RDP, renamed PsExec, AnyDesk remote control and WMI.
  • Defence evasion. Tampering with Defender configuration, custom tooling to kill EDR processes, and log clearing.
  • Exfiltration. MEGASync and Rclone uploads with 7-Zip or WinRAR staging, aimed squarely at finance, HR, medical records and design data.
  • Final deployment. Usually after taking the domain controller, pushed through GPO or bulk scripts with Windows and ESXi executed together.

Because Citrix Bleed-class flaws hijack already-authenticated sessions, MFA provides no backstop on this path - which is why organisations that had two-factor enabled still got breached.

Encryption behavior

Windows build. Salsa20 and AES for symmetric encryption with Curve25519 key agreement and an independent key per file. Command-line arguments select target directories, single files or encryption intensity, letting operators tune speed against thoroughness per environment.

Linux / ESXi build. X25519 ECDH produces a shared secret, SHA-256 derives keys, and AES-128-CTR encrypts content. ESXi-specific routines first enumerate and shut down VMs via vim-cmd - with the option to skip selected VM IDs - so file locks cannot leave encryption incomplete, then process vmdk, vmx and related files under /vmfs/volumes.

Partial encryption. Large files are encrypted partially or intermittently for speed, with the proportion varying by mode. That leaves untouched regions in database files and virtual disks, the primary repair entry point when no decryptor exists - though usability must be measured.

Recovery inhibition. Shadow copies deleted, database and backup services stopped, event logs cleared, and snapshot files handled on the ESXi side.

Multi-layered extortion. Data is exfiltrated before encryption and published in stages if unpaid, with ransom notes printed inside the victim's offices and direct pressure applied to related parties.

Assess before you act

Recoverability assessment

There is no public decryptor for INC Ransom and decryption is not technically viable. We do not pay ransoms and do not negotiate; recovery is entirely technical.

1) Backups, snapshots and shadow copies (first path). INC stops backup services and deletes shadow copies, but offline media, immutable backups, storage array and NAS volume snapshots, hypervisor snapshots and backup servers outside the production domain frequently survive. Because INC usually deploys after taking the domain controller, any backup system authenticating with domain credentials must be assumed reachable - prioritise checking backups with independent credentials.

2) Structural repair of virtual disks and databases (depends on the encryption pattern). INC encrypts large files partially, so vmdk images often retain long intact stretches of guest filesystem metadata and data, allowing partition and metadata repair followed by mounted extraction. SQL Server (MDF/LDF), Oracle (DBF) and MySQL (ibd) files support page-level extraction and logical rebuilds with transaction log roll-forward. Yields depend on whether headers and system tables were overwritten, so sample repairs precede any range estimate.

3) Unencrypted duplicates and log replay. File-server previous versions and recycle bins, endpoint caches, reporting and BI staging databases, ERP/HIS/MES archive exports, mail system copies and application audit logs all support reconstruction or point-in-time replay of critical records.

4) Low-level carving. Where the encryptor wrote new files and deleted originals, source content may remain in unallocated space and can be carved - provided writes to the affected volumes and datastores stop immediately.

5) Handling the data leak. INC runs multi-layered extortion and exfiltrates before encrypting. Establishing the scope and timeline is mandatory because it drives customer notification, regulatory reporting and downstream legal exposure, and cannot be skipped just because files were restored.

There is no guaranteed route for INC Ransom data recovery. We deliver an assessment and an expected range first, then execute, without promising "100% decryption" or guaranteed results.

Our response plan

Hit by INC Ransom ransomware? What to do

  1. Containment and forensic preservation

    Isolate affected Windows servers and ESXi hosts immediately and break storage paths, and prioritise taking the Citrix NetScaler or SonicWall appliances offline while preserving their configuration and session logs - the evidence window for Citrix Bleed-style session hijacking is short. Revoke all active sessions and tokens, since changing passwords alone does not evict an already hijacked session. Do not reboot hosts or re-initialise datastores. Image or take read-only snapshots of domain controllers, backup servers and datastores, and retain three to five .INC files plus INC-README.txt.

  2. Family identification and encryption analysis

    Confirm the family from the .INC extension, the paired INC-README notes and the signature block appended to encrypted files, then separate the Windows build from the Linux/ESXi build - their key systems differ and so do the repair strategies. Also determine whether the sample is a source-derived variant such as Lynx, which shares extensive code but uses independent keys. Measure the coverage and distribution of partial encryption and map intact regions in vmdk and database files.

  3. Recoverability assessment and plan sign-off

    Check backups in layers by credential isolation: offline or immutable backups with independent credentials, then storage and hypervisor snapshots, then domain-joined backup systems - verifying each by actual mount rather than job status. Run sample repairs on core databases and key VMs to assess the viability of page-level rebuilds. Deliver a written assessment covering the recovery path per system, expected ranges, timeline and business sequencing, plus an initial exfiltration scope conclusion.

  4. Recovery execution

    Work on images in a clean environment with original disks and datastores read-only. For ESXi, repair virtual disk structures and extract via mount, landing data on new storage before rebuilding VMs. Databases go through page-level extraction plus transaction log roll-forward, and filesystem data is restored in batches by business directory. Sequence: domain controllers and identity, then core business databases such as HIS, ERP and MES, then file and mail systems, then archives. Each batch gets hash verification, application open tests and business reconciliation.

  5. Attribution, hardening and handover

    Reconstruct the kill chain: unpatched edge appliances and hijacked sessions, which accounts were used for RDP, when the domain controller fell, and the timing and volume of exfiltration. Hardening: upgrade Citrix and SonicWall devices immediately and clear persistent sessions per vendor guidance, reset domain credentials and Kerberos tickets, reduce the management surface of edge appliances, redeploy EDR with tamper protection, segment the ESXi management network with lockdown mode, and rebuild backups around independent credentials, immutable copies and offline media. Close with an incident report, exfiltration conclusion, handover checklist and a scheduled restore drill.

Risk warning

What not to do

  • Do not assume a password change is enough. Citrix Bleed-class flaws hijack already-authenticated sessions, so all active sessions and tokens must be revoked and vendor guidance followed, or the attacker simply returns.
  • Do not reboot ESXi hosts, re-initialise datastores or create new VMs on the original LUN - that ends any chance of vmdk structural repair or carving.
  • Do not use domain administrator credentials to connect to or "check" the backup server; INC usually already controls the domain controller, and this exposes the last surviving backup.
  • Do not delete INC-README.txt / INC-README.html or encrypted samples, and do not clean up suspicious binaries - they are the basis for family and variant identification.
  • Do not bring systems back online before EDR tampering and remote control tools such as AnyDesk are removed; INC routinely keeps that channel for a second encryption run.
  • Do not pay or open contact through the portal; INC runs multi-layered extortion, and payment neither guarantees successful decryption nor prevents the data being reused.

Emergency response

Data already encrypted? Stop and let an engineer look first

We do not pay ransoms and we do not negotiate with attackers. Engineers run a free assessment first, then propose a recovery plan and a firm quote.

Related scenarios

Related industries

Similar families

FAQ

INC Ransom Frequently asked questions

  • Can .INC files be decrypted?

    There is no public decryptor. The Windows build encrypts with Salsa20/AES under Curve25519 key agreement and the Linux/ESXi build derives AES-128-CTR keys via X25519 ECDH, both with per-file keys and no published implementation flaw. Note too that INC's source was sold in 2024, and derivatives such as Lynx share code but use independent keys, so no universal solution exists. Real recovery runs through backups and snapshots, structural repair of virtual disks and databases, unencrypted duplicates and carving.

  • We had MFA enabled - how did INC still get in?

    Because Citrix Bleed (CVE-2023-4966) and similar flaws that INC favours hijack already-authenticated session tokens - the attacker replays a legitimate session without ever passing through login, so MFA cannot intervene. That also shapes the response: resetting passwords achieves nothing, and all active sessions must be revoked, persistent sessions cleared per vendor guidance, and the appliance upgraded immediately. SonicWall SMA series devices were exploited at scale by INC in 2026 and warrant the same priority check.

  • The ransom note came out of our printers - what does that tell us?

    It is one of INC's signature behaviours: once it controls the internal network, it pushes the note to every reachable printer and fax machine to widen internal awareness and increase pressure. Forensically it also means the attacker already holds significant internal privilege and has completed asset discovery - typically the domain controller has fallen and exfiltration is most likely complete. Printed notes should be treated as evidence of full compromise, not an isolated event.

  • Are INC and Lynx the same group, and does it change the response?

    Not the same crew, but the same code lineage. Unit 42 measured roughly 48% overall function overlap between Lynx and INC, rising to 70.8% among common functions and higher still in the Linux/ESXi variant, and the consensus is that Lynx bought or reused the INC source offered for sale in 2024. The practical difference is in identification: encryption behaviour is similar but extensions, notes and key systems differ, so sample analysis must pin down the exact variant or both the recovery plan and the exfiltration investigation can go astray. The recovery techniques themselves are essentially the same.

  • Our backup server was encrypted too - is there any hope?

    Often yes, depending on how the backup estate is layered. INC typically deploys after taking the domain controller, so online backups authenticating with domain credentials are the easiest to wipe out, while appliances with independent credentials, offline tape, offsite copies, storage array or NAS volume snapshots and cloud version history frequently survive. Backup files are also large, and under partial encryption they usually retain substantial intact regions, so structural repair of the backup container and extraction of its contents is worth attempting. The first step is to stop writes and inventory thoroughly, not to rush into rebuilding the backup environment.