What this means for your clients: a stolen GitHub token with write access can read nearly every secret a repository can use, with no malware and no attacker server involved. A branch named ci-sync- plus six random characters appears in a private repository. Its one unsigned commit, ci: sync config, adds a single file, .github/workflows/sync-check.yml, under the name of whoever owned the stolen token. The push runs the workflow, the workflow prints the repository's secrets int

Update, 30 October: CISA added CVE-2026-109618 and CVE-2026-109619 to its Known Exploited Vulnerabilities catalog on October 26. On October 28 it added the Linux kernel flaw CVE-2026-46300, and CISA's KEV entry says that flaw can be chained with both PHP bugs to implant PHP webshells and run commands as root. Dream Research, which calls the kernel flaw Fragnesia, reports a campaign that did exactly that: a custom PHP2Shell exploit plus Fragnesia rooted 25 Palo Alto Networks P

What this means for your clients: a popular tool that connects business apps to AI services briefly published a new version that steals passwords and keys. Any team that installed a compromised 1.107.0 wheel or ran an infected image should treat that environment as compromised, rebuild it, delete any cached copy and change every key that machine could reach. On the morning of October 21, a developer on your client's team pulled the newest LiteLLM to try a new model. The pro

What this means for your clients: if a firewall was open to the internet, attackers may already hold a copy of the passwords stored inside it, so installing the update is only half the job and those passwords need to be changed too. Picture a hypothetical client firewall: patched, ticket closed. Two weeks before the patch, someone downloaded its configuration file, and that copy now sits on a machine you will never see. Inside it are the bind account passwords the firewall us

What this means for your clients: in our sandbox we built the last vulnerable and first fixed PHP 8.5 releases and sent the advisory's own trigger to both under Apache mod_php. On the vulnerable build the Zend memory manager flagged its own heap as corrupted and the worker aborted. On the fixed build the same input was handled cleanly and the worker stayed up. That is strong evidence the update works, not proof against every attack path, so patch now. Update, Oct 24, 2026: We

Update, 30 October: CISA added both PHP2Shell CVEs to its Known Exploited Vulnerabilities catalog on October 26. Dream Research reports a campaign that chained a custom PHP2Shell exploit with the Linux kernel flaw CVE-2026-46300, which it calls Fragnesia, and rooted 25 Palo Alto Networks PAN-OS firewall and management servers. Dream says the attacker targeted the PAN-OS 12.1 and 12.2 branches running PHP 8.0.30, and that the exploit was built for PAN-OS and was not a copy of

What this means for your clients: criminals are now letting AI agents spend a day probing online shops at what NorForge estimates is a few dollars per attempt, and the file they leave behind can look like an ordinary upload form. Take a popular online store with enough traffic to make an attacker's shortlist. Before any agent touches it, a script checks that the store really takes orders. Then a set of AI agents gets one goal: find a way to run code on the site without loggin

What this means for your clients: any client website that switched on a Rank Math PRO 3.0.123 copy downloaded or updated during the affected window on October 15 or 16 should be treated as hacked, and installing the newer version alone does not clean it up. Picture a small agency that looks after a stack of client sites, every one running the same premium WordPress plugin, with updates switched on. On a Thursday afternoon, the agency's scheduled update job pulls the latest bu

A small office runs Nextcloud on one server, and staff edit spreadsheets in the browser without asking how. It works because of one optional app, added from the app store long ago, that downloads a copy of a document server and runs it on the same box. Most days nobody thinks about it. When the machine restarts, there is a short stretch before that bundled server is back, and during it the app's plain PHP file tries to start the server again with each request that comes in. C

Picture an admin opening the Microsoft Entra sign-in logs on a Friday evening and finding a sign-in from an app the company approved long ago. There is no user, no device and no MFA prompt. The source is a hosting provider address, not Microsoft's network. Nobody in the tenant clicked anything, and nobody in the tenant held the credential that signed in. ConnectWise says it identified unauthorized access to application identities associated with its Cloud Backup environment,

A help desk technician starts a remote session to fix a laptop, and the tool does its job. Behind that session sits one appliance that holds the access for everything the help desk can reach. Mandiant says root on it gives an attacker access to every credential the appliance stores or brokers, and both Mandiant and BeyondTrust say this actor deleted logs. The box is the last place to ask what happened. BeyondTrust published advisory BT26-05 on 15 October 2026 for CVE-2026-102

A mail gateway reads every inbound attachment so nobody else has to. Cisco says one crafted attachment passing through an affected gateway can give an unauthenticated sender root. Nobody has to open it. That same root access can tidy the logs on the box, so the box is the last place to ask what happened. Cisco published advisory cisco-sa-esa-afw-zqYJkjxp on 14 October 2026 for CVE-2026-76519, a Critical Secure Email Gateway arbitrary file write bug in the Advanced Malware Pro

Picture someone in finance opening a file that arrived in a chat message. There is no yellow bar, no "this file came from the internet" prompt and no SmartScreen warning. If something runs next, nothing on the screen tells them, and nothing tells you either. Microsoft released CVE-2026-96951 on 13 October 2026 as a Windows Shell Security Feature Bypass Vulnerability, marked Exploited: Yes and Publicly disclosed: No. CISA added it to the Known Exploited Vulnerabilities catalog

Picture a Connect Secure box, fully patched at the time, passing a clean Integrity Checker run after its AD service account has been used to search for LAPS passwords and BitLocker recovery keys. Based on what Volexity found on the device it investigated, that outcome is realistic on this chain. Volexity reported on 13 October 2026 that a previously unknown actor it tracks as UTA0576 chained three bugs in Ivanti Connect Secure to get unauthenticated root. Two are tracked as C

Picture a platform engineer finding a new workflow called Security Audit on the default branch, committed under a longtime maintainer's real name, with no pull request and a run that already finished. The cloud key someone committed years ago and deleted the next morning was in the history it read. StepSecurity reports that on 8 October 2026 two hijacked maintainer accounts pushed a fake "Security Audit" workflow to 345 repositories in two sweeps that each took minutes. It se

Picture an agency developer poking around a client's WordPress server and finding, in a folder that should only hold photos, a file named like an admin login page. The extension is PHP, it lives under wp-content/uploads, and nobody on the team put it there. On 8 October 2026, Patchstack published active exploitation attempts against Bricksforge, a commercial add-on for Bricks Builder, tracked as CVE-2026-85097. It first saw the traffic at 21:47:17 UTC on 7 October (2:47 PM PT

Picture an MSP technician checking the backup server on a Thursday morning because the fans are loud. Task Manager opens, CPU looks normal, and the window closes. An hour later the box is pinned again. The process doing the work is called edge.exe, it lives in a Temp folder, and the service that keeps it alive is called MicrosoftEdgeUpdateSvc. Nobody on the team installed Edge on that server. On 8 October 2026, Huntress reported threat actors chaining two AhsayCBS flaws, CVE-

Chinese state-linked hackers are disguising VPN backdoors as legitimate Windows processes. Here is how to hunt for them in your environment.

CVE-2026-21589 lets unauthenticated attackers read files from eight Atlassian Data Center products. Patch now, block with the WAF regex, and hunt your access logs for the two-dot pattern.

Rejetto HFS CVE-2026-61500 lets attackers forge an admin cookie from a weak Math.random() signing key, then reach remote code execution. Versions 3.0.0 through 3.2.0 are affected; 3.2.1 fixes it. Four hunts to find the activity in your logs.