Ransomware family
DireWolf Ransomware Decryption & Data Recovery
- Active
- High
- No public decryptor
DireWolf (Dire Wolf) is a Go-based ransomware crew that surfaced in May 2025, marked by the .direwolf extension and a HowToRecoveryFiles.txt note. It runs double extortion from a Tor leak site, ships a Windows-only encryptor, and encrypts just the first 1 MB of files larger than 1 MB - which leaves real repair space for databases and virtual disks.
- First seen
- 2025-05
- File extensions
- .direwolf
- Ransom notes
- HowToRecoveryFiles.txt
- Affected platforms
- Windows / Databases
Family profile
- File extensions
- .direwolf
- Ransom notes
- HowToRecoveryFiles.txt
- Contact patterns
- qTox / Tox messenger ID (assigned per victim, not a shared ID)
- Hard-coded live-chat credentials inside the note (roomID / username / password)
- Tor (.onion) negotiation page and leak site
- gofile.io style free file-hosting links showing a sample pack of stolen data
- No mailbox contact
- Aliases / versions
- Dire Wolf、Dire Wolf Ransomware
- First seen
- 2025-05
- Status
- Active
- Operational status
- Actively operating
- Threat level
- High
- Affected platforms
- Windows
- Databases
- Tags
- Leak-site regular
- Active
- Double extortion
- Targets databases
There is no free public decryptor for DireWolf. Neither No More Ransom nor any security vendor has released a tool that recovers .direwolf files.
The encryptor pairs Curve25519 key exchange with ChaCha20 stream encryption, embeds only the operators' public key in the sample, and derives a separate session key per file. Public reverse-engineering work by Trustwave SpiderLabs and AhnLab ASEC has found no exploitable implementation flaw or weak randomness. Without the operators' private key there is no cryptographic path back.
A warning worth repeating: most programs advertised in search results as a "DireWolf decryptor" or ".direwolf decryption tool" are adware bundles, infostealers or second-stage extortion. Running one against original disks generally destroys files that were still repairable. The real recovery space for this family comes from its partial-encryption behaviour, hypervisor snapshots and untouched backups - not from a decryptor.
Latest activity
The DireWolf leak site keeps posting at a high tempo, with roughly 135 named victims across 30+ countries; recent entries include Port of Tanjung Pelepas, RelyComply and Sales Boomerang. Healthcare, professional services and technology lead.
SourcesThree victims posted within 24 hours: Hospital Clinico Universidad de Chile, Erdem Hospital (Turkiye) and Swedish publisher THQ Nordic, with roughly 335 GB claimed from the latter - a clear late-2026 push into healthcare and technology.
SourcesUS organ-donation matching organisation National Kidney Registry was named on the DireWolf leak site, reinforcing the group's 2026 focus on healthcare and health-data organisations.
Sources
Overview
DireWolf - written "Dire Wolf" in most reporting - was first publicly documented by Trustwave SpiderLabs in May 2025. The first batch of victims appeared on its leak site in late May 2025, with some incidents traced back to attacks in April 2025. The operators describe themselves as purely financially motivated with no political stance. The encryptor is written in Go and shipped UPX-packed, and the operation is plainly human-operated: intrude, reconnoitre, exfiltrate, then detonate across the estate at once.
Unlike broad affiliate programmes such as LockBit or Qilin, public analysis points to DireWolf being run directly by a small core team with no open affiliate recruitment. Each build is tailored to one victim - the live-chat roomID, username and password are hard-coded into the sample, which means a manual targeting and packaging step precedes deployment. Demands have reached the USD 500,000 range in single incidents.
From roughly a dozen victims in 2025 to more than 130 named organisations across a dozen-plus countries and territories by September 2026 - trackers count this differently, so figures vary by platform and date - DireWolf has moved from newcomer to leak-site regular in little more than a year. Healthcare, professional services, technology, manufacturing and financial services dominate. Geographically, alongside the United States and Brazil there is a pronounced concentration in Asia-Pacific - Malaysia, Thailand, Singapore and Taiwan.
No public report confirms a targeted campaign against mainland China organisations. For Chinese groups with plants or branches in Southeast Asia, however, the target space overlaps closely with their own overseas footprint.
One gap must be stated honestly: DireWolf's initial access has never been fully confirmed in any public report. Every vendor account of it is inference drawn from victim post-mortems.
How to identify it
Extension. The full original filename is kept and .direwolf is appended, so report.xlsx becomes report.xlsx.direwolf. Names are neither rewritten nor randomised, and the extension is the same for every victim rather than per-target.
Ransom note. HowToRecoveryFiles.txt is dropped in affected directories. The English text states that the network was breached, critical data stolen and files encrypted, offers a limited window before disclosure - public reporting does not record a consistent number of days, so go by the note actually received - and supplies a victim-specific live-chat roomID, username and password plus a qTox ID. A free file-hosting link (typically gofile.io) usually accompanies it, pointing to a sample pack of stolen files as proof.
Host artefacts.
- Mutex Global\direwolfAppMutex, used to prevent re-execution.
- A marker file named runfinish.exe written to the root of the C: drive once encryption completes.
- Application, System, Security and Setup event logs cleared via wevtutil.
- The Windows Event Log service terminated, alongside roughly 75 hard-coded services and 59 processes covering security products, databases and office suites.
- The encryptor self-deletes after execution, so the original sample is frequently gone by the time anyone investigates.
Exclusions (relevant to both attribution and recovery). AppData, Windows, Program Files, $Recycle.Bin and System Volume Information are skipped, as are bootmgr, ntldr and NTUSER.DAT, and the extensions .exe, .dll, .sys, .drv, .iso and .img. The machine therefore still boots and the desktop looks normal - only business files fail to open. That appearance of normality is exactly why many organisations delay reporting and keep using the host, destroying evidence in the process.
Bottom line. The .direwolf extension together with HowToRecoveryFiles.txt is enough to attribute the family. The recovery plan, however, depends on the file-size distribution and which regions were overwritten, so real samples must be analysed before any conclusion is drawn.
Infection vectors
DireWolf's initial access step has never been fully confirmed in public reporting, and this remains the family's largest intelligence gap. The vectors below are vendor inferences from victim post-mortems and should be treated as investigative leads rather than established fact:
- Spear-phishing carrying malicious attachments or links.
- Abuse of internet-facing VPN, RDP and remote-access appliances. No public report ties this family to a named vendor product or a specific CVE, and at least one vendor states plainly that no documented exploitation has been observed.
- Valid account logins to RDP, VPN or remote management tooling, using credentials from password reuse, infostealer logs or dark-web purchases.
- Indirect compromise through third-party providers and the supply chain.
Post-compromise behaviour is far better documented. Operators use living-off-the-land tooling - PowerShell, WMI, PsExec - to map the network, harvest credentials and move laterally. Dwell time runs to several days or more - public analysis gives no consistent figure - exfiltration is concentrated before encryption, and the encryptor is then pushed to many hosts at once during a business trough. Because victim-specific chat credentials are hard-coded into the build, a manual packaging step must occur before deployment, which means defenders have a meaningful detection window between reconnaissance and detonation.
The practical lesson is that this crew does not rely on zero-days. It relies on ordinary weaknesses: VPN and RDP without MFA, unpatched edge appliances, backup consoles sharing domain credentials, and remote management tools that anyone can install. Those are the highest-yield places to look and to harden.
Encryption behavior
Algorithms. Curve25519 key exchange against a hard-coded operator public key, combined with ChaCha20 stream encryption; the shared secret is run through SHA-256 to derive the working key and nonce, and every file gets its own session key. No implementation flaw has been found in public reversing, and without the operators' private key nothing can be reversed.
Partial encryption - the defining technical trait. Files under 1 MB are encrypted in full. Files larger than 1 MB have only their first 1 MB encrypted, with the remainder left untouched. The design is a speed optimisation, but it has a specific consequence: header metadata, database boot pages, virtual disk descriptors and partition structures are destroyed, while the bulk of large files survives intact. Recovery prospects therefore split sharply by file type and size, and quantifying that split is the first step in any recovery plan.
Recovery inhibition.
- vssadmin delete shadows /all /quiet removes every volume shadow copy.
- wbadmin deletes Windows Backup images and bcdedit disables recovery options.
- Roughly 75 hard-coded services are terminated, covering security products such as Sophos, Symantec and Qihoo 360 as well as database and backup services including SQLServerAgent, MSExchangeIS, VeeamTransportSvc and BackupExecJobEngine - the point is to release exclusively-locked database and mail files so they can be encrypted.
- About 59 processes are killed, event logs are cleared, and the binary self-deletes when finished.
Platform. As of September 2026 public analysis has only observed a 64-bit Windows encryptor, and at least one vendor states explicitly that no Linux or ESXi build exists. That is a point-in-time judgement: if the crew ships a Linux or ESXi encryptor, the assessment below must be revisited immediately. Damage in virtualised estates therefore usually comes from files being encrypted inside guest VMs rather than the hypervisor datastore being locked wholesale - a distinction with real value during assessment, and one to verify immediately.
Double extortion. Data is exfiltrated before encryption and a sample pack is linked from the note as proof. Non-payment leads to staged publication on the Tor leak site - public reporting gives no consistent interval before full release - alongside direct pressure on the victim's customers and partners. Volumes vary by incident and must be established from actual traffic and logs rather than estimated from precedent.
Assess before you act
Recoverability assessment
There is no shortcut to decrypting DireWolf, but "cannot be decrypted" is not the same as "everything is lost". We do not pay ransoms and do not negotiate on a client's behalf; our work is technical recovery and forensics. Assess in this order:
1) Free decryptor: none exists. Neither No More Ransom nor any vendor has published a tool for .direwolf, and no exploitable weakness has been disclosed in the Curve25519 + ChaCha20 implementation. This path is currently zero and should not form the basis of a plan.
2) Structural repair of large files - the main opportunity with this family. Because files over 1 MB have only their first 1 MB encrypted, SQL Server MDF/NDF, Oracle DBF, MySQL ibd, PST/OST mail stores, vmdk/vhdx virtual disks, large archives and media files usually retain intact bodies; what is destroyed is the header, boot pages and some metadata. Databases can be addressed with page-level extraction and logical rebuilds; virtual disks can often be mounted and their contents extracted after partition table and filesystem metadata repair. The achievable ratio depends on which critical structures happened to sit inside that first 1 MB, so it has to be measured on the real files rather than predicted from the extension. The converse must be stated plainly: files under 1 MB - most Word, Excel and PowerPoint documents, source code and configuration files - are fully encrypted, and this path does nothing for them.
3) Backups, snapshots and shadow copies. Shadow copies are normally gone, but everything else deserves checking: offline and offsite backups, storage-layer snapshots on NAS, SAN or gateways, untouched copies on the backup server, and cloud version history. One point deserves emphasis - because DireWolf currently ships only a Windows encryptor, hypervisor snapshots and the host datastore are frequently untouched. VM-level backups and ESXi or Hyper-V snapshots are often the highest-yield route in these incidents and should be verified rather than assumed lost. Never reattach backup media to a network that has not been cleaned.
4) Unencrypted copies and log replay. The encryptor skips AppData, Windows and Program Files, so Outlook caches and some application data in default endpoint paths may survive completely. File-server recycle bins, BI and reporting staging databases, ERP archive exports, database transaction logs and application audit logs can all support reconstruction or point-in-time replay.
5) Low-level carving. Where the build writes a new encrypted file and deletes the original, source data may remain in unallocated clusters and can be recovered by raw sector scanning. This requires stopping all writes to the affected volumes immediately.
We commit to a verifiable assessment and a clearly bounded recovery scope. Claims of complete decryption or certain restoration have no technical basis; the volume that can actually be retrieved is established by testing real samples.
Our response plan
Hit by DireWolf ransomware? What to do
Containment and forensic preservation
Cut affected hosts off from production networks and storage paths and disable suspect VPN, RDP and remote management accounts - but do not power off or reboot. DireWolf clears the Application, System, Security and Setup event logs and self-deletes afterwards, so on-host logging is largely worthless; the priority is exporting evidence that lives elsewhere: EDR and antivirus console records, firewall and VPN gateway logs, Active Directory logs, and backup platform job history. Image or snapshot the domain controller, backup server, hypervisor management hosts and core database servers first, and preserve three to five encrypted files spanning both sides of the 1 MB boundary plus the original HowToRecoveryFiles.txt.
Family identification and encryption profiling
Confirm the family from the .direwolf extension, the structure of HowToRecoveryFiles.txt, the Global\direwolfAppMutex mutex and the runfinish.exe marker at the root of C:. Then perform the step that matters most for this family: sample and quantify the encrypted regions by file class - verify that large files really do have only their first 1 MB overwritten, identify exactly which structures were destroyed (database boot pages, virtual disk descriptors, partition tables, file headers), and measure how the affected data splits above and below the 1 MB boundary. This output determines how much can be repaired and which systems qualify; recovery discussed without it is guesswork.
Recoverability and exfiltration impact assessment
Two tracks run in parallel. Recovery: inventory hypervisor snapshots, storage-layer snapshots, offline and offsite backups and untouched copies on the backup server, then run sample repairs against the critical databases and virtual disks to prove out page-level extraction and structural rebuilding. Exfiltration: DireWolf is a double-extortion operation, so the timing, scope and volume of data theft must be established in parallel - it drives regulatory notification, contractual obligations and customer disclosure. The deliverable is a written assessment stating which systems roll back from backup, which go to structural repair and which must be written off, with expected recovery ranges, timelines and a business restoration order. Execution follows sign-off, not before.
Recovery execution and business verification
All work happens on images or copies with originals kept read-only. Restore in business priority order: domain controllers and identity first, then core databases such as ERP, MES and finance, then file services and mail. In virtualised environments, roll back from host-side snapshots or VM-level backups wherever possible, and only fall back to structural repair of vmdk/vhdx files when nothing else survives. Database work combines page-level extraction with transaction log replay to pull data as close as possible to the pre-incident point in time. After each batch, run integrity checks and business-side verification - reconciliation, report comparison, application start-up and key process testing - recorded in a traceable recovery manifest.
Attribution, hardening and handover
Because DireWolf's entry point is still unsettled, attribution has to rebuild the kill chain from off-host telemetry: anomalous VPN and RDP logins by geography or hour, unpatched edge appliances, unauthorised remote management tool installs, where domain administrator credentials leaked, and the timing and destination of outbound transfers. Remove residual persistence, rogue accounts, scheduled tasks and GPO backdoors; reset credentials domain-wide and enforce MFA on VPN, RDP and backup consoles; decouple the backup console from domain credentials; constrain PsExec and WMI lateral movement and block unauthorised remote tooling; rebuild backups to a 3-2-1 design with immutable copies. Close with an incident report and a formal handover checklist.
Risk warning
What not to do
- Do not keep using an affected host just because it still boots. DireWolf skips system directories so the machine looks normal, but every boot and write overwrites unallocated space and directly erodes the chance of low-level carving.
- Do not reboot or power off affected hosts - losing memory-resident processes, connections and traces makes forensics, already hampered by cleared event logs, considerably harder.
- Do not download or run any "DireWolf decryptor" found online. No free decryptor exists for this family; such programs are usually adware, infostealers or second-stage extortion, and running one against original disks can destroy large files that were still repairable.
- Do not delete HowToRecoveryFiles.txt or the encrypted samples, and do not rush antivirus "virus file" cleanup - they are the only basis for build identification and for measuring which regions were encrypted.
- Do not format, reinstall or rebuild RAID sets and storage pools, and never create new VMs on the original LUN or re-initialise a datastore; doing so closes off structural repair and carving at the same time.
- Do not contact the operators through the qTox ID or chat credentials in the note, and do not pay. Payment buys no assurance of a working key and does not prevent stolen data from being published or resold.
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
Domain Controller Compromise and Estate-Wide Encryption
A compromised domain controller hands the attacker a legitimate administrator identity, allowing an encryptor to be pushed to every host at once through Group Policy or remote execution. This page covers how such incidents present, the correct order for Active Directory recovery, and how to decide between cleanup and full rebuild.
Database Encrypted by Ransomware
When database files are encrypted, every business system that depends on them stops at once. This page explains how we triage an encrypted database, how recoverability is assessed, and when file repair, backup-plus-log restore, or rebuild is the right path.
Backups Deleted or Destroyed
Modern ransomware follows a fixed sequence: destroy the backups, then encrypt the data — deleting shadow copies, encrypting repositories, disabling jobs, and exploiting backup software flaws to steal credentials. This page covers what can still be inventoried once backups fail, why replication propagates encrypted files off-site, and what offline and immutable copies are really worth.
File Servers and NAS Encrypted by Ransomware
When shared folders on a file server or NAS are encrypted, drawings, contracts, archives, quotations and design sources all become unusable at once — and mapped drives spread the impact to every endpoint. This page covers how to gauge spread, what shadow copies and snapshots realistically offer, and how to sequence recovery by business value.
Related industries
Manufacturing Ransomware Response and Recovery
Ransomware in manufacturing hits information systems and production cadence at the same time: with ERP down there are no orders, with MES down there is no schedule, and an encrypted drawing library takes the process documentation for an entire product line with it. This page covers the asset profile, recovery priorities and targeted defences.
Healthcare Ransomware Response and Recovery
When a hospital is hit, registration, consultation, orders, billing, laboratory and imaging fail at the same moment and care falls back to paper. This page covers the healthcare threat picture, a recovery priority built around clinical continuity, and the handling of patient data and compliance obligations.
Financial Services Ransomware Response and Recovery
Financial and quasi-financial institutions face far stricter requirements on data integrity, transaction continuity and regulatory reporting than most sectors, so one ransomware event hits availability, customer trust and compliance simultaneously. This page covers the threat profile, a recovery approach centred on transactional consistency, and hardening priorities.
Similar families
- No public decryptor
Qilin
Qilin (formerly Agenda) is a Rust-rewritten cross-platform RaaS operation focused on VMware ESXi and Linux estates. It has ranked as the world's most active ransomware group for several consecutive quarters since 2025, with confirmed victims among electronics manufacturers in Taiwan and Hong Kong.
- Some versions decryptable
Akira
Akira is a ransomware-as-a-service operation that emerged in March 2023, breaking in through VPNs without MFA and edge-device flaws, then encrypting Windows estates and VMware ESXi clusters under double extortion. CISA's November 2025 advisory update calls it an imminent threat to critical infrastructure.
- No public decryptor
Lynx
Lynx is a RaaS operation that emerged in mid-2024 and shares extensive code with INC Ransom. It ships encryptors for Windows, Linux and ESXi through an affiliate panel with an 80/20 split and had passed 400 claimed victims by 2026. No public decryptor exists.
FAQ
DireWolf Frequently asked questions
Can .direwolf files be decrypted?
No free decryptor exists. DireWolf uses Curve25519 key exchange with ChaCha20 and a separate session key per file, and public reversing has found no exploitable flaw, so without the operators' private key there is nothing to reverse. That does not mean the data is gone. Because files over 1 MB have only their first 1 MB encrypted, large databases, virtual disks and mail stores usually retain intact bodies that structural repair can recover a substantial part of, and hypervisor snapshots and VM-level backups are often untouched. Get an assessment before concluding anything - and never trial an unknown "decryptor" against original disks.
The note sets a deadline before disclosure - should we respond?
That deadline is negotiation pressure, not a technical clock, and it should not drive decisions. DireWolf exfiltrates before encrypting, the operators control the publication schedule, and public reporting records no consistent interval. Paying does not remove the risk either - the operators still hold a copy that can be resold or recycled by other crews. We do not pay ransoms or negotiate. That window should be spent on isolation and evidence preservation, verifying whether backups and hypervisor snapshots are usable, and establishing the scope and timeline of the exfiltration so internal notification and compliance obligations can start.
Our SQL Server or Oracle database was encrypted to .direwolf - can it be repaired?
Often yes, and this is where DireWolf differs from most families. Production MDF/NDF, DBF and ibd files are typically far larger than 1 MB, and only the first 1 MB is encrypted, so the damage is concentrated in headers and boot pages while the vast majority of data pages remain original. Page-level extraction and logical rebuilding can recover table data, and surviving transaction logs support point-in-time replay. The achievable ratio depends on which critical structures - system tables, allocation bitmaps - happened to fall inside that first 1 MB, so a range can only be given after testing the real files. Also check the backup platform and hypervisor snapshots first: services such as SQLSERVERAGENT were killed before encryption, but host-level snapshots frequently survive.
Files inside our VMs were encrypted - is the ESXi host gone too?
Most likely not, but verify immediately. Public analysis has only observed a Windows encryptor for DireWolf, with no confirmed Linux or ESXi build, so damage normally occurs inside guest VMs rather than the hypervisor datastore being locked wholesale. That means ESXi or Hyper-V snapshots, VM-level backups and storage array snapshots are very likely still usable, and they are frequently the highest-yield recovery route. Confirm whether those snapshots exist and whether they were deleted before doing anything else - and do not create new VMs on the original LUN or re-initialise the datastore.
The machine still boots and only files fail to open - did encryption not finish?
No. DireWolf deliberately skips AppData, Windows, Program Files and $Recycle.Bin, and leaves .exe, .dll and .sys files alone along with bootmgr, ntldr and NTUSER.DAT, specifically so the system stays bootable and the victim can read the note and negotiate. Encryption has finished - a runfinish.exe marker is usually left at the root of C: to prove it. That appearance of normality is the dangerous part: organisations keep working on the machine, overwrite unallocated space, and spend the carving option without realising it. Stop using the host and isolate it now.
What should we do in the first hour after discovery?
Four things. First, isolate: break production network and storage paths and disable VPN, RDP and remote management accounts - but do not power off or reboot. Second, capture logs: DireWolf clears four event log channels and self-deletes, so on-host logging is unreliable; immediately export records from EDR, the firewall, the VPN gateway, Active Directory and the backup platform, all of which live off the host. Third, preserve samples: keep three to five encrypted files spanning both sides of the 1 MB boundary, plus the original HowToRecoveryFiles.txt. Fourth, establish whether hypervisor snapshots and backups still exist and whether they were touched - that is the highest-yield recovery route for this family. We run 24/7 emergency response and can usually return an initial family assessment and recovery path within an hour of remote access.
Sources
- Dire Wolf Strikes: New Ransomware Group Targeting Global Sectors — Trustwave SpiderLabs / LevelBlue
- Dire Wolf Ransomware: Threat Combining Data Encryption and Leak Extortion — AhnLab ASEC
- Dire Wolf — Halcyon Ransomware Threat Group Profile
- Deep Dive Analysis into Dire Wolf Ransomware — TTPs and IOCs — Protos Labs
- direwolf — 受害者与泄露站活动统计 — Ransomware.live
- Decryption Tools(未收录 .direwolf 解密器)— No More Ransom
External links are provided for reference only. The content is published by third parties and does not represent our position.
Updated