Luna Moth, which the FBI tracks as the Silent Ransom Group, steals data from law firms and other professional services firms by talking employees into opening a remote session, and then extorts them without ever deploying ransomware. The group needs no malware, no exploit and no stolen password: a phone call, a legitimate remote support tool and an employee who believes they are helping IT are enough. This profile summarises what the FBI and reporting security vendors have published about the group, maps its behaviour to MITRE ATT&CK, and sets out the detection, prevention and response controls that address it. It is written for security teams, IT managers and practice leaders in law firms, insurers, financial and healthcare organisations. For the wider technique, see the Callback Phishing Guide and the Vishing and Smishing Guide.
Key takeaways
- Data theft, not encryption. The FBI's May 2026 FLASH describes data theft and extortion without encryption-based ransomware, so backups alone do not resolve an incident.
- The lure moved from fake invoices to fake IT. Early campaigns used subscription renewal emails with a phone number; by 2025 the group was calling staff while posing as their own IT department.
- Someone may turn up in person. The FBI reports that when remote access fails, the group has sent a person to the office to plug in a storage device, claiming to image or back up the computer.
- Every tool is legitimate. Zoho Assist, Quick Assist, AnyDesk, RustDesk, Syncro, Splashtop and Atera for access, and WinSCP or a renamed Rclone for exfiltration, so antivirus rarely fires.
- The best controls are process controls. A published rule for how IT contacts staff, a block on unapproved remote support tools, and visitor verification stop most of the attack chain.
Who they are (names and aliases)
The FBI's Private Industry Notification 20250523-001 (23 May 2025) and FLASH-20260526-01 (26 May 2026) name the group Silent Ransom Group (SRG) and list the aliases Luna Moth, Chatty Spider and UNC3753. Luna Moth is the name used by Sygnia and by Palo Alto Networks Unit 42, which published a detailed incident response account in November 2022. Both FBI products state that the group has been active since 2022.
| Attribute | What the sources say | Source |
|---|---|---|
| Names | Silent Ransom Group (SRG); Luna Moth; Chatty Spider; UNC3753 | FBI PIN 2025, FBI FLASH 2026 |
| Active since | At least 2022 | FBI FLASH 2026 |
| Model | Data theft and extortion without encryption | FBI FLASH 2026; Unit 42 |
| Main targets | US-based law firms since spring 2023; also insurance, finance, medical and healthcare organisations | FBI PIN 2025, FBI FLASH 2026 |
| Earlier targets | Small and medium legal businesses, then larger retail organisations (mid-May to late October 2022 campaign) | Unit 42 |
| Leak site | A public site used inconsistently; the 2026 FLASH names business-data-leaks[.]com | FBI PIN 2025, FBI FLASH 2026 |
| Lineage | Unit 42 notes that AdvIntel links the actor to the Silent Ransom Group and to Conti, and to the earlier BazarCall callback campaigns; Unit 42 said it could not confirm the Conti connection | Unit 42 |
The lineage claim is reported by one vendor through another and was not confirmed by Unit 42, and neither FBI product mentions it. Treat it as context, not as a basis for defence. For the earlier Conti operation, see the Conti Ransomware Profile Guide.
Targeting and motivation
The motive is financial. The FBI notes that law firms are targeted likely because of the highly sensitive nature of legal data, and that most victims are law firms or firms with similar naming conventions. Client confidentiality obligations, privileged material and reputational exposure give the attacker leverage that encrypted servers do not. Unit 42 observed demands ranging from 2 to 78 BTC, set according to the victim's researched revenue, with discounts of about 25 percent for prompt payment and a unique wallet per victim. EclecticIQ research, as reported by BleepingComputer in May 2025, put later demands at between 1 million and 8 million US dollars depending on victim size.
Payment does not buy deletion. Unit 42 reported that some victims who paid received no further response, and that the actor escalated to threatening the victim's own customers when victims stayed silent. The FBI adds that the group calls employees and clients to pressure the victim into negotiation. For the general decision framework, see the Data Extortion Without Encryption Guide.
How the attacks work
Phase 1: subscription callback lures (from 2022)
Unit 42 described personalised emails, sent from legitimate free email services, claiming the recipient had been charged for a subscription, usually under 1,000 US dollars. The email contained no link and no malware: only a PDF invoice impersonating an online class platform or a health club membership service, a confirmation number and a phone number to cancel. The FBI notes the small fees were chosen because they are less likely to raise immediate suspicion. A caller who rang to cancel was queued for a live agent, who sent a Zoho Assist invitation, took control of keyboard and mouse, blanked the screen and searched the computer and connected shares. Where the victim had administrator rights, Syncro was installed for persistence. Exfiltration happened during the call or hours to weeks later.
Phase 2: fake IT help desk (from early 2025)
By April 2025, according to the FBI, the group was impersonating the victim's own IT department, and the 2026 FLASH describes this as its current method. The FBI describes direct calls to employees and phishing emails urging staff to call the actor, followed by a request to grant a remote desktop session. EclecticIQ research reported by BleepingComputer found at least 37 domains registered by March 2025, most imitating help desk or support portals of US law and financial firms with patterns such as [company_name]-helpdesk.com, from which victims were told to install a remote monitoring and management (RMM) tool.
Phase 3: in-person visits
The FBI first reported in May 2025 that a person posing as IT support was sent to the firm to access a computer and insert a storage device. The 2026 FLASH adds that this happens when remote access fails, and that the visitor claims the device is needed to image the computer or create a backup. Exfiltration then uses external hard drives or USB drives as well as WinSCP, Rclone, Google Drive and Microsoft OneDrive. This is the step most organisations have no control for, because it relies on reception and office staff rather than technology.
Tactics and techniques (MITRE ATT&CK)
MITRE ATT&CK v20 cites the Unit 42 and Sygnia Luna Moth reports in its description of Spearphishing Voice. The mapping below is drawn from the FBI and Unit 42 descriptions.
| Stage | Technique | Observed behaviour |
|---|---|---|
| Resource development | T1583.001 Domains | Look-alike help desk domains; earlier .xyz domains (Unit 42) |
| Initial access | T1566.004 Spearphishing Voice | Callback invoices and calls posing as IT |
| Initial access | T1200 Hardware Additions | In-person visitor inserts a storage device |
| Stealth | T1684.001 Impersonation | Posing as subscription support or the firm's IT staff |
| Execution | T1204 User Execution | Employee installs or launches the remote support tool |
| Command and control | T1219.002 Remote Desktop Software | Zoho Assist, Quick Assist, AnyDesk, RustDesk, Syncro, Splashtop, Atera |
| Collection | T1005 Data from Local System; T1039 Data from Network Shared Drive | Manual search of the laptop and mapped shares |
| Stealth | T1036.003 Rename Legitimate Utilities | Hidden or renamed Rclone |
| Exfiltration | T1048 Exfiltration Over Alternative Protocol; T1567.002 Exfiltration to Cloud Storage; T1052.001 Exfiltration over USB | WinSCP over SFTP, Rclone to Google Drive or OneDrive, USB drives |
Privilege escalation is minimal according to the FBI: the actor works with the rights of the employee it deceived. That is why least privilege on file shares and document management systems limits the damage more than endpoint hardening does.
How they abuse identities and access
- The employee's session is the credential: no password is stolen. The attacker inherits whatever the user can reach through mapped drives, synced folders and open applications.
- IT's identity is the lure: the attack works because staff cannot tell a real IT call from a fake one. A defined way for IT to prove who it is removes the ambiguity. See the Help Desk Identity Verification Policy Template.
- Persistence through a tool account: where Syncro or another RMM agent is installed, the attacker holds an agent enrolled in its own tenant, which survives a password reset. Removal of the agent, not just the user's credential, is required.
- Personal cloud accounts as the exit: exfiltration to Google Drive or OneDrive uses accounts the attacker controls, so tenant restrictions and controls on personal cloud storage matter.
Detection and defence
Controls mapped to Luna Moth behaviour
| Behaviour | Preventive control | Detection |
|---|---|---|
| Calls claiming to be IT | Published rule: IT never asks staff to install tools or open sessions on an unsolicited call; staff call back on the internal number | A reporting route for suspicious calls, logged and reviewed for clusters |
| Remote support tool installation | Application control that allows only the firm's approved remote support tool and blocks portable versions, as CISA advisory AA23-025A recommends | Alert on first execution of any RMM binary not on the approved list, including Quick Assist where it is not used |
| Look-alike help desk domains | Block newly registered domains and domains containing your firm name plus "helpdesk" or "support" | DNS and proxy logs for such patterns; domain registration monitoring |
| In-person "IT" visitor | Verify every visitor with IT before access; copy visitor ID as the FBI recommends; disable external drive use on sensitive computers | Device control logs for new USB mass storage on fee earner machines |
| WinSCP and Rclone | Block unapproved file transfer tools; restrict outbound SSH where it is not needed (the FBI suggests blocking port 22 where feasible) | Executables with Rclone or WinSCP characteristics under other names; SFTP or bulk HTTPS uploads from user devices |
| Bulk access to shares | Least privilege on matter folders and the document management system | Unusual file read volume per user and per session |
What to monitor
- Endpoint: installation or first run of remote support software, process names that do not match the file's original name metadata (renamed Rclone), and WinSCP Portable launched from user folders.
- Network: connections to RMM vendor cloud relays that your firm does not use, outbound SFTP to external addresses, and large uploads to personal cloud storage.
- Identity and data: spikes in document management downloads or share enumeration by one user during a call window.
- People: reports of unexpected IT calls, voicemails or visitors; the FBI lists unsolicited calls from people claiming to work in IT as an indicator in itself.
The FBI cautions that use of these remote access tools should not be treated as malicious without evidence of actor control. Detection should drive a quick verification with the user, not an automatic incident. See the RMM Tool Security Guide for allowlisting design.
Rolling out the controls
- Week 1: publish the IT contact rule to all staff and reception, with the internal number to call back, and brief office managers on the in-person variant.
- Weeks 1 to 2: inventory remote support tools in use and pick one approved tool; run application control in audit mode to see what else executes.
- Weeks 3 to 4: enforce the block on unapproved RMM tools and portable file transfer tools, starting with fee earners and finance; keep an exception process with a named approver.
- Month 2: enforce device control on sensitive machines, add DNS blocking for look-alike help desk domains, and add detections for renamed utilities and bulk uploads.
- Month 3: run a simulated fake IT call and an unannounced visitor test; measure reporting, not just failure.
- Rollback: if a block disrupts client work, move that rule back to audit mode for the affected group only and record the exception; do not remove the control for everyone.
Small firms without a security team can still apply the first, third and fourth steps through their managed IT provider, and should ask that provider how it will identify itself when it calls. Larger firms should integrate the reporting route with the SOC so that clusters of calls are visible across offices.
Responding to an incident
- Contain: end the remote session, isolate the device, remove any RMM agent enrolled during the call and block its tenant or relay where possible.
- Scope: review what the user could reach and what was read or transferred; WinSCP and Rclone configuration files and logs on the device, proxy logs and cloud storage logs show destinations. See the Data Exfiltration Scoping Guide.
- Preserve: the FBI asks for the ransom note, the phone numbers or email accounts used, voicemails and call transcripts, the original phishing email or callback message, cryptocurrency wallet details, a description of the data taken and, for visitors, surveillance video and photos.
- Report: the FBI asks victims to contact their local FBI field office and to report to IC3; sharing is voluntary and must follow applicable law. Consider client, regulator and privilege implications early. See the Extortion Email Playbook and the Legal Privilege in Incident Response Guide.
Metrics that show it is working
- Unapproved RMM executions: count per month of remote support tools blocked or detected outside the approved list; should fall after enforcement and then stay near zero.
- Report rate: percentage of staff who report a simulated fake IT call; target above 70 percent.
- Visitor verification: percentage of unannounced "IT" visitors stopped at reception in tests; target 100 percent.
- Device control coverage: percentage of sensitive endpoints blocking unapproved USB storage.
- Time to contain: minutes from a staff report to termination of the remote session.
Framework mapping
| Framework | Reference | Relevance |
|---|---|---|
| NIST SP 800-53 Rev. 5 | AT-2 Literacy Training and Awareness | Training on callback and fake IT calls |
| NIST SP 800-53 Rev. 5 | CM-7 Least Functionality; AC-17 Remote Access | Blocking unapproved remote support and file transfer tools |
| NIST SP 800-53 Rev. 5 | PE-2 Physical Access Authorizations; PE-8 Visitor Access Records; MP-7 Media Use | Visitor verification and removable media control |
| NIST SP 800-53 Rev. 5 | AC-6 Least Privilege; SI-4 System Monitoring | Limiting reachable data and spotting bulk transfer |
| NIST CSF 2.0 | PR.AT-01, PR.AA-05, DE.CM-01, DE.CM-03 | Awareness, access permissions, network and personnel activity monitoring |
Common failure patterns
- "We do not let users install software": portable remote support and file transfer tools run without administrator rights, and Quick Assist ships with Windows.
- No rule for how IT calls: staff cannot refuse a caller if they have never been told what a real IT request looks like.
- Reception not in scope: security awareness reaches fee earners but not front desk and facilities staff, who meet the in-person visitor.
- Treating it as a ransomware case: restoring from backup changes nothing; the issue is stolen client data and notification duties.
- Removing the user's password but not the agent: an RMM agent enrolled to the attacker's account keeps working after a credential reset.
- Paying for deletion: Unit 42 observed victims who paid and heard nothing further, with no proof the data was destroyed.
Practitioner checklist
- Publish and repeat a rule for how IT contacts staff and how staff verify a caller.
- Allow one approved remote support tool; block other RMM tools and their portable versions.
- Block or alert on WinSCP, Rclone and renamed copies on user devices, and restrict outbound SSH where it is not needed.
- Require reception to verify any IT visitor with the IT team before granting access, and record visitor ID.
- Enforce device control on computers that handle sensitive client data.
- Block look-alike help desk domains for your firm name and newly registered domains.
- Apply least privilege to matter folders and monitor bulk reads per user.
- Prepare an extortion response plan covering evidence the FBI requests, legal privilege and client notification.
Standards and references
- FBI: FLASH-20260526-01, Silent Ransom Group Impersonating IT Personnel through Social Engineering
- FBI: Private Industry Notification 20250523-001, Silent Ransom Group Targeting Law Firms
- Palo Alto Networks Unit 42: Luna Moth callback phishing extortion campaign
- CISA: AA23-025A Protecting Against Malicious Use of Remote Monitoring and Management Software
- MITRE ATT&CK: T1566.004 Spearphishing Voice
- MITRE ATT&CK: T1219.002 Remote Desktop Software
- BleepingComputer: Luna Moth extortion hackers pose as IT help desks (EclecticIQ research)
Related NHI Mgmt Group resources: Callback Phishing Guide · Vishing and Smishing Guide · Law Firm Identity and Data Security Guide · Vishing and Help Desk Social Engineering Awareness Guide · Data Extortion Without Encryption Guide