The Ridge View is Ridgeview Advisors’ weekly roundup of what happened in the MSP channel. For the week of November 9, 2026, five stories stood out: a day-long Microsoft 365 outage, a worm spreading through ScreenConnect and an urgent N-central hotfix, new GTIA research on the MFA gap, the first standalone AI liability policy for small businesses, and Sherweb opening billing data to AI agents.
This week at a glance
- Microsoft 365 outage: Exchange Online, Outlook, Teams, Copilot and admin portals degraded from November 12 into November 13.
- SCKit worm: Rogue ScreenConnect clients landed on 288 endpoints across 48 organizations in about 90 minutes, then 205 more hosts in a second wave.
- N-central 2026.4 HF2: N-able told on-premises customers to patch now and keep servers off the internet until they do.
- GTIA research: 42 percent of providers have MFA on all critical systems for most of their customers.
- Cowbell AI liability: A standalone policy for AI agent incidents goes on sale January 1, 2027.
- Sherweb MCP server: AI assistants can now read Sherweb subscription, billing and renewal data.
What happened in the Microsoft 365 outage?
Microsoft reported degraded Exchange Online and Microsoft 365 service starting November 12. The outage reached most of the tools a client’s day runs on, and recovery stretched into a second day.
- Affected services: Outlook access, mail delivery, Teams features, Microsoft 365 Copilot, admin portals and parts of Defender for Office 365.
- Recovery: By November 13 Microsoft reported improved availability and shrinking email queues.
- Lingering impact: Some users still saw intermittent access and delayed mail after service improved.
By the numbers: Two days of degraded service across mail, chat, AI and the admin portals MSPs use to respond.
Sources: BleepingComputer (opens in a new tab), ChannelPro (opens in a new tab)
What is the SCKit worm and how did it spread through ScreenConnect?
Huntress first saw SCKit activity at 17:47 UTC on November 11. All 288 endpoints in the first wave, across 48 organizations, were hit within about 90 minutes, roughly a third of them in the first ten minutes. A second wave began November 13 on new infrastructure and reached 205 hosts across 41 organizations.
Huntress says the activity “seems to be related” to the third-party RMM+ extension, versions 1.0.31 and 1.0.32. RMM+ checks requests against a shared access key, and if an admin never set one, the key is blank and an empty value passes. Because the extension bypasses ScreenConnect user accounts, roles and MFA, commands reached endpoints with no matching login activity.
How the attack worked:
- Entry: An operator used Run Command on cloud-hosted panels to execute commands as SYSTEM across many endpoints.
- Rogue client: A second ScreenConnect client was installed silently, followed by Defender exclusions and AppLocker rules.
- Relay hijack: A scheduled task called TrapWatch runs every 10 seconds, stops the legitimate client and takes over its relay connection.
- Cookie theft: The kit reads browser cookie stores for the technician’s ScreenConnect session cookie and replays it to find other reachable sessions.
ConnectWise published an advisory at 22:45 UTC on November 11 and disabled five third-party extensions across about 400 partner environments: Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link and RMM+ Password Link. ConnectWise also urges technician clients to run 26.6.5 or later, which fixes CVE-2026-84869, a critical file-transfer flaw.
“Do not connect to it through ScreenConnect.”
Huntress, on handling suspect hosts
What Huntress recommends:
- Review ScreenConnect Run Command and Toolbox history for unexpected mass activity.
- Check endpoints for an unauthorized second ScreenConnect client, and investigate from a clean device using WinRM or another out-of-band tool.
- Isolate affected hosts, then remove the rogue client service, the TrapWatch task, the C:\ProgramData\RelayTrap folder and the RelayTrapLiteAlert Run key.
- End active ScreenConnect web sessions, rotate ScreenConnect credentials and revoke exposed browser and identity provider sessions.
- Disable the listed extensions on on-premises deployments.
Source: Huntress (opens in a new tab)
What did N-able tell N-central customers to do?
In a separate security response, N-able released N-central 2026.4 HF2 to address the highest-risk vulnerabilities disclosed this month.
- On-premises customers: Install HF2 as soon as possible and keep servers off the internet until the upgrade is complete.
- All customers: Update Agents and Probes.
- Hosted customers: N-able has already upgraded hosted servers and returned them to service.
- What is fixed: Every identified vulnerability an unauthenticated user could exploit.
- What remains: Issues that require an authenticated account, often with elevated privileges, will be addressed in later hotfixes.
Source: ChannelPro (opens in a new tab)
What did GTIA’s research find about cybersecurity and the MFA gap?
GTIA published two pieces of research on November 9: its Channel Trends in Cybersecurity study and its Q3 2026 Quarterly Cybersecurity Tracker, which surveyed 309 North American IT service providers. Demand for security services is rising, while baseline controls at customer sites lag behind.
Demand and revenue:
- 76 percent plan to add new cybersecurity solutions in the next 12 months, up from 65 percent in 2025.
- Nearly 60 percent plan to add AI-related cybersecurity solutions.
- 87 percent report winning a client because they offer cybersecurity services.
- 21 percent report losing a client to a cybersecurity incident in the past year.
Controls in place at customers:
- 42 percent deploy MFA on all critical systems for more than half their customers.
- 35 percent deploy immutable, air-gapped backups broadly.
- For each of the six controls GTIA measured, fewer than half of providers deploy it to more than half their customers.
- 51 percent name AI-generated attacks as a leading customer concern.
“Providers can turn that concern into a clear action plan.”
Carolyn April, Vice President, Research and Market Intelligence, GTIA
Source: GTIA (opens in a new tab)
What does Cowbell’s new AI liability insurance cover?
Cowbell announced on November 10 a standalone AI liability policy for U.S. small and midsize businesses that deploy AI agents in areas like customer service, payments or core operations. It becomes available January 1, 2027, sold only through wholesale brokers on a non-admitted basis.
- Trigger: An “AI Incident,” such as an AI tool exceeding its authorized scope, acting without required approval, exposing data or producing inaccurate or unsafe output.
- First-party coverage: Incident response, output correction and notification, direct transaction losses, runaway AI usage charges, and LLMjacking charges from stolen credentials or API keys.
- Third-party coverage: Defense, damages and settlements for claims of algorithmic discrimination, copyright or trademark infringement, bodily injury or property damage.
- Underwriting: Three AI-specific factors assess a business’s AI observability, autonomy and governance.
“AI liability is no longer only about people using AI.”
Simon Hughes, Chief Commercial Officer, Cowbell
Source: Cowbell (opens in a new tab)
What is Sherweb’s MCP server and what data does it expose?
Sherweb launched the Cumulus MCP Server on November 6, giving approved AI tools read-only access to its Partner Portal. What a user can see follows their existing Partner Portal permissions.
- Data available: Organizations (tenants), subscriptions, the product catalogue, billing, audit logs and renewal details.
- Supported tools: Claude Desktop and Claude Code, ChatGPT, Cursor, Microsoft Copilot Studio, and VS Code with GitHub Copilot.
- Setup: Create an API credential under Security > APIs in the Partner Portal, add the server to an MCP host, authenticate and confirm the tools appear.
- Example uses: Looking up a customer’s subscriptions, checking upcoming renewals and querying billing in plain language.
- What is next: Write access is planned for a future release.
“For now, integrations should be designed around available read-only tools.”
The Sherweb Team
Source: Sherweb (opens in a new tab)
Subscribe on YouTube: youtube.com/@howtomsp (opens in a new tab) Connect on LinkedIn: linkedin.com/in/andrewpmoore (opens in a new tab)


