These news items are automatically aggregated from industry sources and are not individually reviewed. Any inaccuracies are unintentional —
let us know and we'll correct or remove it.
CISA added CVE-2026-21962, a maximum severity CVSS 10.0 flaw in Oracle HTTP Server and the WebLogic Server Proxy Plug-in, to its Known Exploited Vulnerabilities catalog on August 24, giving federal agencies just 72 hours to remediate. Oracle shipped the fix back in January 2026, but researchers now say exploitation began around the same time, with attack patterns detected by CloudSEK in early February and reports of a China-linked actor using it against government targets since. A patch applied on schedule does not tell you whether an attacker already got in before you applied it.
GitHub Actions has logged 13 separate incidents across nine different days in the first 17 days of August 2026, including a 10-hour outage on August 6 that failed 71 percent of workflow runs at peak, and a nearly 8-hour platform-wide incident on August 17 that alone consumed close to a year’s worth of Actions’ allowable downtime budget. Ninety-day uptime for Actions now sits at 99.33 percent, an order of magnitude worse than the 99.99 percent GitHub maintains for core Git operations. For teams whose deploys, tests and release gates run entirely through Actions, the pattern is now frequent enough to plan around rather than treat as a one-off.
GitLab shipped an out-of-band patch on August 17 for CVE-2026-19478, a CVSS 9.4 unauthenticated code injection flaw in its GraphQL API that lets a remote attacker modify or delete public projects and user data with no login required. Public reporting confirmed active exploitation from August 20. Every self-managed GitLab Community and Enterprise Edition instance from 18.2 through 19.2 that has not yet applied 18.11.11, 19.0.8, 19.1.6 or 19.2.4 is exposed to an attack that needs nothing more than network access to the instance.
Latvia’s Road Traffic Safety Directorate (CSDD) confirmed that a cyberattack over the weekend of August 8-9 exposed payment records for more than 1.2 million people and 200,000 organisations, dating back to 2008. CERT.LV traced the intrusion to a vulnerability in a CSDD system reachable from the internet. Officials have since resigned, a criminal investigation is underway, and the president has asked whether the breach amounts to a threat against critical state infrastructure. For any organisation running a public-facing system holding registry-scale personal data, this is what an unpatched exposure looks like at the far end of the consequence scale.
Palo Alto Networks’ Unit 42 uncovered a Chinese-speaking threat actor, tracked as knaithe, who wired DeepSeek into an open-source tool called Hermes Agent and let it scan, research and exploit targets almost entirely on its own, hitting over 460 organisations across eight CVEs including Citrix NetScaler, Langflow and n8n. Unit 42 confirmed the actor tried the same setup with Claude and OpenAI models first, and both providers’ safety controls blocked the attempt. The operation was only discovered because the actor misconfigured Hermes Agent and exposed the entire operational workspace, credentials and target lists included.
DigitalOcean opened public preview of Spot GPU Droplets, offering NVIDIA HGX B300 and AMD Instinct MI350X and MI355X capacity at a rate locked below on-demand pricing for the life of the instance, aimed at short-lived batch training, inference and rendering workloads. The preview is live only in DigitalOcean’s US datacenters, with no European region included at launch. For European teams already running GPU workloads on DigitalOcean or evaluating it against Hetzner, AWS and sovereign cloud alternatives, the gap is worth planning around rather than waiting out.
Attackers hijacked a maintainer account on crates.io and published poisoned releases of arrayref, internment and append-only-vec on August 20, each pointing to a typosquatted dependency whose build script downloaded and ran a payload the moment a project compiled. The Rust Security Response Team pulled the malicious versions within 86 to 107 minutes, but arrayref alone has 245 million all time downloads and 403 crates listing it as a direct dependency, meaning the exposure window did not need to be long to matter. No application code had to run, and no developer had to click anything, compiling was enough.
AWS added a fourth Availability Zone, eu-west-2d, to its Europe (London) Region on 19 August, bringing new EC2 Trainium and P6 accelerated instance capacity alongside general purpose compute at standard London pricing. The stated driver is demand for AI and ML capacity, but the practical value for most UK and European customers is the extra fault isolation boundary it creates for architectures still built on the region’s original three zones. Capacity announcements like this only pay off for organisations that actually redesign around the new zone rather than simply noting it exists.
Microsoft disclosed CVE-2026-69836, a maximum-severity deserialization flaw in Entra ID that let an unauthenticated attacker execute code with no user interaction and no privileges required, and confirmed it was already being exploited before the fix shipped. The vulnerability was mitigated server-side, so no customer patch exists to install, but Microsoft has not said who exploited it, when, or how widely. For any European organisation whose sign-ins, Conditional Access and app registrations run through Entra ID, that silence is the whole problem.
CERT Polska issued Alert 145/2026 warning that CVE-2026-73570, an unauthenticated OS command injection flaw in Zimbra Collaboration Suite’s SNMP notification handling, is under active exploitation. Shadowserver counts more than 12,100 exposed Zimbra servers online, with the largest single regional concentration in Europe. A fix has existed since version 10.1.20 shipped on 20 July, more than a month before the exploitation warning, which means every vulnerable server still online is running on borrowed time rather than missing information.
This site uses cookies. By continuing to use this website, you agree to their use.
We’ll help you resolve your infrastructure challenges
Our team of experts is ready to help you with your infrastructure challenges. We’ll give you honest and personal treatment. Get in touch to learn more.
