In short
MeshCentral is a free, open-source, self-hosted web server for remote device management. Its MeshAgent client gives administrators remote desktop, terminal and file access to Windows, Linux, macOS and FreeBSD machines. Attackers abuse it as a legitimate-looking backdoor, so defenders should allow only approved MeshCentral servers.
Key takeaways
- What it is
- A free, Apache-2.0-licensed remote management web server built on Node.js. Its agent, MeshAgent, runs on Windows, Linux, macOS and FreeBSD.
- How it works
- Each managed device runs MeshAgent, which connects out to the MeshCentral server over a WebSocket. Administrators then open desktop, terminal and file sessions through the browser.
- Why it matters
- Kaspersky, Microsoft and MITRE ATT&CK all document intruders using MeshCentral or MeshAgent for remote access, from espionage to ransomware.
- What Odin Research saw
- 232 MeshCentral relay records behind TacticalRMM servers in Monograph 12, many on fresh disposable domains. Hostile use was not shown, so all stayed on watchlists.
- What to do first
- Keep an allowlist of approved RMM servers, block unapproved MeshAgent binaries even when signed, and alert on renamed agents.
What is MeshCentral?
MeshCentral is a free, open-source web server for remote monitoring and management (RMM). An organization hosts it itself and installs the MeshAgent client on devices. Administrators then get remote desktop, terminal and file access from a browser. Attackers abuse the same features as a backdoor that looks like normal IT work.
The project's website calls it an open-source, multi-platform, self-hosted site for remote device management, built on Node.js and available on npm. The GitHub repository, under the account Ylianst, is licensed under Apache-2.0. It had about 7,700 stars and 1,200 forks on 21 November 2026.
MeshCentral also sits inside other products. TacticalRMM, a popular open-source RMM, uses it for remote desktop and shell sessions. That makes MeshCentral traffic common in managed service provider (MSP) networks, and hard to judge on sight.
How MeshCentral works
A MeshCentral deployment has two parts:
- The server. A Node.js web application that administrators reach from a browser. The website says the server runs on Windows, Linux, macOS and FreeBSD.
- MeshAgent. A small agent on each managed device. It connects out to the server and waits for commands. Kaspersky's analysis describes a MeshAgent configuration file that sets the server address, reached over a WebSocket.
Because the agent dials out, the managed device needs no open inbound port. An administrator who clicks desktop, terminal or files in the browser starts a session that the server relays to the agent.
Inside TacticalRMM, the documentation describes three hostnames. MeshCentral runs on the mesh. name. Nginx serves it over HTTPS on port 443 and forwards to MeshCentral on local port 4430. TacticalRMM's "Take Control" remote desktop feature uses MeshCentral.
Odin Research's Monograph 12 adds the wire view. Remote shells and desktops pass through a WebSocket to /meshrelay.ashx, and a signed MeshAgent.exe keeps that WebSocket open over port 443. To a firewall, it looks like ordinary HTTPS.
How attackers use MeshCentral
RMM tools are attractive to intruders. They are signed, trusted and built for remote control. CISA, NSA and MS-ISAC warned in advisory AA23-025A that threat actors can abuse any legitimate RMM software for persistence and command and control. Several vendors have documented MeshCentral in particular:
- Espionage. On 7 October 2024, Kaspersky reported that the Awaken Likho group replaced its UltraVNC module with MeshAgent. A scheduled task named to look like a Microsoft Edge update launched the agent and connected it to the attackers' MeshCentral server. Targets included Russian government agencies, their contractors and industrial firms.
- Ransomware. On 24 September 2026, Microsoft Threat Intelligence reported that MeshAgent is one of the most frequently observed tools of Storm-2570. The actor used it as a bridge from early hands-on access to discovery, credential theft and ransomware. Renamed MeshAgent binaries were pushed to other hosts with PsExec. In one intrusion, MeshAgent came before the deployment of Qilin.
- Data theft. MITRE ATT&CK lists MeshCentral under Remote Access Tools (T1219). It records that ShinyHunters used MeshCentral and ConnectWise to gain initial access and run administrative commands.
The common thread is living off the land: a legitimate tool, often renamed, doing what it was built to do for the wrong operator.
What Odin Research observed
Monograph 12 studied MeshCentral as part of the TacticalRMM stack, in a chain that runs from traffic gates to remote management to relays.
- 232 relay records. The report counted 232 records on
mesh.hostnames serving the MeshCentral relay, on port 443 or 4430, next to 405 agent-enrollment API records. - Disposable domains. An estimated 338 TacticalRMM servers sat on newly registered
.online,.xyz,.topor dynamic DNS domains. Legitimate MSPs usually run TacticalRMM on corporate domains that are years old. - Signed and persistent. Once a machine is enrolled, the signed MeshAgent gives an operator PowerShell and desktop access that blends into HTTPS. That is why the domain behind the connection matters more than the software.
- No hostile use shown. Odin Research did not see these servers used against a victim. All stayed on watchlists, not block lists.
The same report found FRP relays and Keitaro traffic gates. The access chain across these tools is an analytical model, not a link between specific servers.
How to detect unapproved MeshCentral use
- Agent processes. Alert when
MeshAgent.exeruns on a device that has no approved RMM, or connects to a server outside your list. - Renamed binaries. Microsoft suggests hunting process, file and registry events for
meshagent,meshagent64andmeshcentral, which also catches renamed files and services. - Relay traffic. A WebSocket upgrade to
/meshrelay.ashxon a domain you do not manage is a strong lead. This needs TLS inspection or proxy logs. - Click-to-install chains. A browser visit to a fresh domain followed within minutes by
msiexec.exe, PowerShell or an RMM agent launch deserves review. - Persistence. Look for services and scheduled tasks that start an RMM agent under a misleading name.
Monograph 12 publishes four rules. They cover WebSocket upgrades to /meshrelay.ashx (Suricata), a Sigma rule for unapproved TacticalRMM or MeshCentral agents, a YARA rule for TacticalRMM bootstrap files and a Sentinel KQL hunt for browser-to-installer chains. Fill in your approved MSP domains before use. All rules are in the Defense Kit.
How to defend against MeshCentral abuse
- Allowlist RMM by server. Permit MeshCentral and other RMM traffic only to documented corporate or MSP names, at DNS and at the proxy. Block the rest by default.
- Use application control. MITRE and CISA both recommend allowlisting. Block unapproved
MeshAgent.exewith WDAC or AppLocker even when the file carries a valid signature. - Limit RMM to approved paths. CISA advises using authorized RMM only inside the network or through approved VPN or VDI access.
- Respond to a find. Microsoft advises resetting the passwords of accounts used to install an unapproved RMM, and digging deeper if a SYSTEM-level account was used. See External Remote Services.
MeshCentral compared with other remote access tools
| Tool | Type | Abuse documented by | Detection angle |
|---|---|---|---|
| MeshCentral | Self-hosted open-source RMM | Kaspersky (Awaken Likho), Microsoft (Storm-2570), MITRE (ShinyHunters) | Unapproved server, renamed MeshAgent, /meshrelay.ashx |
| TacticalRMM | Open-source RMM that embeds MeshCentral | Odin Research watchlist (Monograph 12) | Disposable rmm., api. and mesh. domains |
| AnyDesk and ScreenConnect | Commercial remote desktop | CISA advisory AA23-025A | Portable executables outside IT installs |
| Remote access trojans | Malware | Many vendors | Unsigned or packed binaries, malware C2 |
Frequently asked questions about MeshCentral
What is MeshCentral?
MeshCentral is a free, open-source, self-hosted web server for remote device management. Its MeshAgent client gives administrators remote desktop, terminal and file access to managed computers.
Is MeshCentral malware?
No. MeshCentral is legitimate, Apache-2.0-licensed software. Attackers abuse it as a remote access backdoor, so an unapproved MeshAgent on your network should be treated as a possible intrusion.
Is MeshCentral safe to use?
Yes, when you host it yourself, keep it patched, protect logins with MFA and allow agents to talk only to your own server. Risk comes from agents that connect to servers you do not control.
What is MeshAgent?
MeshAgent is the client that runs on each managed device. It connects out to a MeshCentral server, often over a WebSocket, and relays desktop, terminal and file sessions.
Who made MeshCentral?
MeshCentral is an open-source project hosted on GitHub under the account Ylianst and distributed on npm.
Which threat actors use MeshCentral?
Kaspersky reported that Awaken Likho switched to MeshAgent in 2024. Microsoft linked MeshAgent to the ransomware affiliate Storm-2570. MITRE ATT&CK records ShinyHunters using MeshCentral for initial access.
How do I detect MeshCentral on my network?
Hunt for meshagent, meshagent64 and meshcentral in process, file and registry data, agents connecting to unknown servers, and WebSocket traffic to /meshrelay.ashx. Odin Research publishes Suricata, Sigma, YARA and KQL rules for these signals.
MeshCentral vs TacticalRMM: what is the difference?
MeshCentral is a remote management server on its own. TacticalRMM is a fuller RMM platform that uses MeshCentral for its remote desktop and shell features, on a separate mesh subdomain.
Sources
- MeshCentral project website. meshcentral.com/ (checked 21 November 2026)
- MeshCentral repository, GitHub. github.com/Ylianst/MeshCentral (checked 21 November 2026)
- Kaspersky Securelist, Awaken Likho APT campaign with a new implant, 7 October 2024. securelist.com/awaken-likho-apt-new-implant-campaign/114101/ (checked 21 November 2026)
- Microsoft Threat Intelligence, "Beyond the ransomware: Tracking Storm-2570's consistent tradecraft across deployments", 24 September 2026. www.microsoft.com/en-us/security/blog/2026/09/24/beyond-ransomware-tracking-storm-2570-consistent-tradecraft-across-deployments/ (checked 21 November 2026)
- MITRE ATT&CK, T1219 Remote Access Tools. attack.mitre.org/techniques/T1219/ (checked 21 November 2026)
- CISA, NSA and MS-ISAC, Advisory AA23-025A, Protecting Against Malicious Use of Remote Monitoring and Management Software. www.cisa.gov/news-events/cybersecurity-advisories/aa23-025a (checked 21 November 2026)
- Tactical RMM documentation, How It All Works. docs.tacticalrmm.com/howitallworks/ (checked 21 November 2026)
Related tools and terms
What is noVNC?
noVNC is an open-source VNC client that runs in a web browser. Phishing kits abuse it to show victims a real browser hosted by the attacker.
Tunneling toolWhat is Ligolo-ng?
Ligolo-ng is an open-source tunneling tool that builds a network tunnel from a compromised host through a TUN interface.
Tunneling toolWhat is FRP?
Fast Reverse Proxy (FRP) is an open-source tool that exposes services behind NAT through a public relay server.
Tunneling toolWhat is ReverseSSH?
ReverseSSH is an open-source tool that gives an operator an SSH shell through a connection the target opens outward.
Shell managerWhat is Platypus?
Platypus is an open-source Go tool for managing many reverse-shell sessions at once.
Device controllerWhat is Hak5 Cloud C2?
Hak5 Cloud C2 is a controller for Hak5 implant devices, such as drop boxes, that call home over the internet.
Odin Research
Get the rules, not just the definition
Every Odin Research report ships with verified indicators and detection rules you can load today.