Act now:Chrome Zero-Day CVE-2026-125350: XML Use-After-Free. Everyone running Chrome on Windows, macOS, Linux or Android below the fixed builds, and admins of fleets where updates are pinned, deferred or waiting on a relaunch.
Act now:Able2Shell. Every managed service provider and IT team running N-central on-premises, and hosted customers too, because all Agents and Probes must be upgraded to HF3.
Check exposure:bee-js 13.3.0: npm Release Drops Wallet Stealer and Backdoor. Anyone whose lockfiles, caches, mirrors or CI logs show @ethersphere/bee-js 13.3.0, including through swarm-cli, bee-dashboard or Swarm Desktop builds that resolved a fresh ^13 range on or after November 9, 2026.
Check exposure:Qubole AWS Role Abuse: Customer Secrets and S3 Stolen. Current and former Qubole customers on AWS whose accounts still contain an IAM role trusting account 805246085872, or IAM access keys created for a Qubole integration.
Do you run any of these? IBM Aspera Faspex, OpenBSD OpenSSH, SignalWire FreeSwitch, GPUStack GPUStack. They are being exploited in the wild (CISA KEV). Patch or mitigate if you do; CISA’s federal deadline is 2026-11-12.
PyPI: 4 critical advisories (fastapi-code-generator, gpustack). Check with pip-audit (or: pip show <package>).
Go: 2 critical advisories (github.com/pocket-id/pocket-id/backend, github.com/kubeflow/pipelines). Check with govulncheck ./....
None of it touches your stack? Then there is nothing to do today. That is a result, not a failure to read carefully.
Google says an exploit exists in the wild and the fix is only starting to roll out, so update and relaunch Chrome today and check whether your other Chromium browsers have a fix yet.
Must act: Everyone running Chrome on Windows, macOS, Linux or Android below the fixed builds, and admins of fleets where updates are pinned, deferred or waiting on a relaunch.
Can relax: Users already on Chrome 157.0.8103.50 or later, or Extended Stable 156.0.8078.96 or later, after a relaunch. Headless and server software that does not embed Chromium is not involved.
Open chrome://settings/help, let the update finish, and click Relaunch. The fix does not apply until the browser restarts.
In managed fleets, report Chrome versions and flag every endpoint below 157.0.8103.50, or below 156.0.8078.96 on Extended Stable.
Do not wait for the staged rollout on high-risk machines. Push the update through your management tool and force a relaunch with the RelaunchNotification policy.
This is a small package and the bad version is already gone from npm, so most teams have nothing to do, but any machine or CI job that installed 13.3.0 with lifecycle scripts enabled should be treated as compromised.
Must act: Anyone whose lockfiles, caches, mirrors or CI logs show @ethersphere/bee-js 13.3.0, including through swarm-cli, bee-dashboard or Swarm Desktop builds that resolved a fresh ^13 range on or after November 9, 2026.
Can relax: Projects that do not depend on bee-js, and projects whose lockfile pinned 13.2.0 or earlier with no fresh resolution during the window, are not affected.
Search lockfiles, internal mirrors, caches and CI logs for @ethersphere/bee-js 13.3.0, including transitive installs.
Pin a reviewed release with npm install @ethersphere/bee-js@13.2.0 --save-exact and regenerate lockfiles that resolved 13.3.0.
If 13.3.0 installed with lifecycle scripts enabled, isolate the host, preserve evidence and rebuild it from a trusted image, because removing the package does not remove the persistence.
Upgrade on-premises N-central to 2026.4 HF3, which permanently blocks the exploited Legacy API, then upgrade every Agent and Probe and hunt for compromise from the HF2 window.
Must act: Every managed service provider and IT team running N-central on-premises, and hosted customers too, because all Agents and Probes must be upgraded to HF3.
Can relax: Organizations with no N-central server and no MSP using one. N-able has named only N-central, and N-able.app services behind N-able Login remain online.
Upgrade on-premises N-central to 2026.4.1.39 (HF3); it installs directly over 2025.4 and later builds, including HF1 and HF2.
Upgrade every Agent and Probe to 2026.4.1 HF3, on hosted and on-premises alike, even those already on HF2, then review the Inactive Assets tab.
Until HF3 is installed, confirm the Legacy API is disabled and keep the server off the internet or behind a VPN.
If any of your AWS accounts has an IAM role that trusts Qubole's account, treat it as possibly used by an attacker and check CloudTrail today; if none does, this incident does not touch you, but the same weakness applies to every vendor role you have.
Must act: Current and former Qubole customers on AWS whose accounts still contain an IAM role trusting account 805246085872, or IAM access keys created for a Qubole integration.
Can relax: AWS users who never integrated Qubole and have no role trusting that account. AWS itself was not compromised.
List IAM roles in every account whose trust policy names account 805246085872, and find any access keys created for Qubole.
Search CloudTrail from October 1, 2026 for AssumeRole events from that account that came from non-AWS IP addresses.
If you no longer use Qubole, delete the roles and keys. If you do, narrow the trust policy to a specific principal and add an aws:SourceIp condition.