[ctx]ctxseccontext security: attacks on the harness and context layer of AI agents, plus exploited-now news, hunting and playbooks.

← Playbooks

Your RMM vendor says take the server offline: run the MSP outage and compromise check (N-central Able2Shell, Nov 2026)

For when an RMM platform has public, unpatched exploit chains and the vendor pulls hosted service and tells on-prem customers to shut down (N-able N-central, Nov 9, 2026). Built from N-able's status updates and Huntress's guidance.

For: MSP owner / operations, SOC, IT admin

1. Take it offline and prove it2. Keep clients running without the console3. Look for use before shutdown4. Restore only on the vendor's word5. Reduce the next blast radius
  1. Step 1 of 5

    Take it offline and prove it

    Do what the vendor says, then confirm from outside that the console no longer answers.

    • Shut down or firewall on-prem N-central servers; block inbound at the edge, not just the service
    • Check from an external host that the web console and agent ports no longer respond
    • Note the exact time it went dark for your incident timeline
  2. Step 2 of 5

    Keep clients running without the console

    Use the vendor's fallback paths instead of standing up ad-hoc remote access tools.

    • N-able Login / n-able.app for asset view, patching, scripting and Take Control
    • Tell clients what is paused (monitoring alerts, scheduled scripts, patch windows)
    • Don't open RDP or install unvetted remote tools as a workaround
  3. Step 3 of 5

    Look for use before shutdown

    Exploit code was public before the shutdown; check from Nov 8 onward.

    • Console logins and admin account creation you can't explain
    • New or edited scripts, automation policies and scheduled tasks pushed to devices
    • Remote sessions and file transfers outside normal hours
    • On endpoints: new services, remote access tools or local admins created by the RMM agent
  4. Step 4 of 5

    Restore only on the vendor's word

    Apply the hotfix first, then bring the server back on a restricted network.

    • Install the vendor's validated hotfix and confirm the build number
    • Restrict console access to VPN or allowlisted IPs before reopening
    • Rotate console admin passwords, API keys and any credentials stored in the RMM
  5. Step 5 of 5

    Reduce the next blast radius

    An RMM server can push code to every client at once.

    • Require MFA and IP allowlists for the console
    • Alert on mass script deployment and new admin accounts
    • Keep an out-of-band way to reach clients that doesn't depend on one RMM

Original content by ctxsec. Printable checklist: use your browser's print view.