Patched Tuesday, Mining Monero by Thursday

N-able N-central's 4EverAble2Shell chain (Axis CVE-2026-90436) put miners on HF2 servers; here's how to check and lock down your clients' N-central.

On Tuesday, 10 November 2026, N-able shipped an emergency hotfix for its N-central remote management server. By Thursday lunchtime UTC, 19 servers running that exact hotfix were mining Monero for a stranger, Huntback reports. The chain is called 4EverAble2Shell. The one setting it needs sits on a single console page.

TL;DR

  • On 12 November 2026, an operator used the 4EverAble2Shell chain to plant XMRig miners on at least 21 N-central servers, 19 on Hotfix 2, per Huntback; N-able reported no compromise, per BleepingComputer.
  • Exposed: on-premises N-central with the Legacy API turned on and reachable from the internet, a non-default setting, N-able said via Huntress. Hosted N-central is not affected.
  • Turn Legacy APIs off under Administration > Mail and Network Settings > Network Security, per N-able, then hunt for /tmp/.xvd and xvd.service.
  • No fix exists yet, and Huntback warns that root on N-central could let attackers move on to the client environments it manages.

What is 4EverAble2Shell?

4EverAble2Shell is an unauthenticated chain to root on N-able N-central, the remote monitoring and management (RMM) console MSPs use to run client machines, published by researcher Sean Bales on 12 November 2026. Bales's write-up says it was verified against build 2026.4.1.26 HF2 and adds that "Legacy API needs to be enabled."

It has two stages, as Huntback describes them:

  1. Get in. A crafted SOAP request (an old XML-based API format) to /dms/services/ServerUI makes the bundled Apache Axis library build a class that runs a shell command. That is CVE-2026-90436, rated 9.8 (CVSS), per The Hacker News. Commands run as the nable service account.
  2. Become root. A symlink fools a privileged N-central helper into handing nable a root-owned script, which the attacker edits, runs as root, then restores. Previdian tracks this step as VULN-2026-0009; the chain has no CVE.

Think of a side door N-able bricked up in 2025 but left a switch to reopen. Its 2025.4 release notes warn that re-enabling these APIs "will expose the N-central server to potential malicious exploits."

Is 4EverAble2Shell an N-able Hotfix 2 bypass?

Yes, by the researcher's testing: HF2 fixed seven Able2Shell CVEs on 10 November 2026, and 4EverAble2Shell works on HF2 when the Legacy API is on. Huntress reports that servers on HF2 with the Legacy API off are not exposed to this chain. The earlier dump is covered in how the first Able2Shell hotfix was beaten in a day.

Is the N-central zero-day being exploited?

Yes. Huntback says the newest N-central zero-day was used against production servers within hours. Its timeline, all UTC on 12 November 2026:

  • 03:58 PoC added to the public repository; updated at 07:48 to note the Legacy API requirement.
  • 08:30 First probe from 157.20.241[.]16.
  • 09:16 First full chain: root, then the miner installer.
  • 12:50 The mining pool lists 21 rigs named nc_<victim-ip>, all active.

Huntback found 19 of the 21 on 2026.4.1.26, most behind MSP-branded login pages across seven countries, but does not know whether each had the Legacy API on. N-able said it had seen no signs of compromise from the new flaws, according to BleepingComputer, so the count rests on Huntback alone. The chain is not in CISA's KEV catalog, Tenable notes.

A miner sounds harmless next to ransomware, but this is an MSP RMM vulnerability: root on N-central sits one step from every client endpoint, and the next operator may not settle for half your CPU.

How do I check if my clients' N-central server was hit?

Run checks 1 and 3–6 in a root shell on the on-premises N-central server, yours or a client's; check 2 uses the console and an outside workstation. Indicators come from Huntback.

  1. Preserve evidence. Snapshot running processes, then copy logs and persistence locations and hash them. Pass: $D.sha256 lists every file.

    D=/root/ir-$(hostname)-$(date +%F); sudo mkdir -p "$D"
    ps -eo pid,ppid,user,comm,args --forest | sudo tee "$D/ps.txt" >/dev/null
    sudo cp -a /var/log /var/opt/n-central/tmp /var/opt/n-central/flag /etc/systemd/system /etc/cron.d /var/spool/cron "$D"/
    sudo cp -a /tmp/.xvd "$D"/tmp-xvd; sudo cp -a /var/tmp/.xvd "$D"/vartmp-xvd
    sudo find "$D" -type f -exec sha256sum {} + | sudo tee "$D.sha256"
    
  2. Check the Legacy API. Open Administration > Mail and Network Settings > Network Security > LEGACY APIS. Then, from outside the network, see whether the SOAP endpoint answers. Pass: the setting is off. Any HTTP code other than 000 means the endpoint is internet-reachable.

    curl -sk -m 10 -o /dev/null -w '%{http_code}\n' https://<n-central-host>/dms/services/ServerUI
    
  3. Search the copied logs for the exploit class, the weaponised user agent and the attacker IP. Pass: no output.

    sudo grep -rniE 'TerminalLineSettings|python-requests/2\.31\.0|157\.20\.241\.16' "$D"/log
    
  4. Look for miner files and persistence, the root-escalation leftovers, and Huntback's hash for the escalation script (usually piped straight to Python, so a miss is weak). Pass: the cron listing shows only entries you recognise and every other command prints nothing.

    sudo ls -la /tmp/.xvd /var/tmp/.xvd /etc/systemd/system/xvd.service /etc/cron.d/xvd 2>/dev/null
    sudo ls -la /etc/cron* /var/spool/cron/*; sudo grep -rn 'xvkeep' /etc/cron* /var/spool/cron/
    sudo find /var/opt/n-central/tmp -name 'serverchild-*' -ls
    sudo find /var/opt/n-central/flag -name flag_default.json -type l -ls
    sudo find /tmp /var/tmp /var/opt/n-central -xdev -type f -exec sha256sum {} + | grep -i fc4c01fd11bc301cf312830460cfdccc3af9dcfe50c910b86d6a1117bc76c747
    
  5. Check running processes for the miner's flags or its .xvd folder. This shows processes still running only. Pass: no output.

    ps -eo pid,ppid,user,comm,args --forest | grep -E -- '--rig-id nc_|--donate-level 1|\.xvd/' | grep -v grep
    
  6. Check live connections to the payload host and both mining pools, then grep the client's firewall export for the same IPs. Pass: no output from either.

    sudo ss -tnp | grep -E '157\.20\.241\.16|139\.99\.69\.109|79\.76\.51\.56'
    grep -rnE '157\.20\.241\.16|139\.99\.69\.109|79\.76\.51\.56' <firewall-log-export>
    

How to disable N-central Legacy API and clean up

There is no fix for 4EverAble2Shell yet, so turning off the Legacy API is the fix for now, and any hit above means a rebuild.

  1. Preserve evidence before anything changes. Pass: every line reads OK.

    sudo sha256sum -c "$D.sha256"
    
  2. Turn off LEGACY APIS at Administration > Mail and Network Settings > Network Security, as N-able advises. Pass: after a page reload the setting shows off.

  3. Block the indicators at the client's perimeter firewall: deny outbound traffic to 157.20.241.16, 139.99.69.109 and 79.76.51.56. Pass: run on the N-central server, this prints blocked.

    timeout 5 bash -c '</dev/tcp/139.99.69.109/10128' && echo OPEN || echo blocked
    
  4. Rebuild any server with a hit. Huntback says to treat it as fully compromised and rebuild rather than clean up. Export Administration > User Management > Users to CSV (N-able) first, so you can compare accounts afterwards. Pass: checks 4 and 5 meet their pass conditions on the new server.

  5. Rotate credentials on a rebuilt server and in the client environments it manages, as Huntback advises. Pass: a login with an old password fails.

  6. Rerun checks 2 to 6. Pass: each check meets its own pass condition.

Agent2Shell, Bales's second release, claims SYSTEM on managed Windows agents. Huntress has not confirmed whether it also needs the Legacy API, so its earlier advice to cut the server's internet access still stands.

Where KinoSec fits

The setting that decided this attack doesn't show up in a version number. KinoSec's autonomous web app pentest finds which of your clients' N-central consoles and endpoints answer from the internet, tests what an outsider can reach, and writes up remediation you can hand the client. With autonomous web app and API pentesting for your clients, you can rerun that outside-in test whenever a new PoC drops, as remote-access tools keep drawing fire (the BeyondTrust remote support zero-day is another). KinoSec does not read N-central's logs or processes, so checks 3–6 stay with your team.

Frequently Asked Questions

Is my N-central server vulnerable to 4EverAble2Shell if I installed Hotfix 2?

It can be. Bales verified 4EverAble2Shell against build 2026.4.1.26 HF2 with the Legacy API on, and Huntback found 19 of 21 mined servers on that build. Huntress reports that HF2 servers with the Legacy API off are not exposed to this chain.

How do I check if my clients' N-central server was hit?

On the server, run sudo ls -la /tmp/.xvd /etc/systemd/system/xvd.service. Then run pgrep -af -- '--rig-id nc_' for a running miner. Finally run sudo grep -rn TerminalLineSettings "$D"/log on copied logs. Huntback says to treat any hit as a full compromise.

Is hosted N-central affected?

No. N-able said hosted environments are not vulnerable to this issue, as quoted by Huntress. The exposure is on-premises N-central with the Legacy API turned on and reachable from the internet.

Is the N-central cryptominer campaign the only risk?

No. Huntback warns that attackers could use compromised N-central servers against the customer environments they manage, because the chain yields root.