Vai al contenuto principale

#vulnerability

Centralizing Cybersecurity Vulnerability Data: A Solution to Fragmented Risk Assessment and Prioritization#cybersecurity

Altro...

Centralizing Cybersecurity Vulnerability Data: A Solution to Fragmented Risk Assessment and Prioritization#cybersecurity
Centralizing Cybersecurity Vulnerability Data: A Solution to Fragmented Risk Assessment and Prioritization

#cybersecurity #software #coding #community #engineering #development #inclusive #vulnerability #risk #centralization

0

Caricamento...

0
1

Caricamento...

GitLab Patches Maximum-Severity RCE Flaw as Exploit Code Goes Public

Read Time:3 Minute, 17 Second

A public proof-of-concept exploit is now circulating for a maximum-severity GitLab vulner

Altro...

Read Time:3 Minute, 17 Second

A public proof-of-concept exploit is now circulating for a maximum-severity GitLab vulnerability, giving organizations running self-managed instances a hard deadline to patch. The flaw, tracked as CVE-2026-97229, carries a CVSS score of 10.0 and allows an attacker to execute commands on a vulnerable GitLab server through a bug in its CI/CD regular-expression compiler.

How the Exploit Chain WorksResearcher Jonah Burgess published the working exploit code, which targets both GitLab Community Edition and Enterprise Edition. GitLab’s own advisory describes the underlying issue as an integer overflow in the CI/CD regular-expression compiler — the component responsible for validating pipeline configuration syntax.

The attack path runs through GitLab’s CI lint endpoint, a feature meant to check CI/CD configuration files for errors before they run. According to the published analysis, an attacker crafts a pipeline configuration containing a regular expression paired with a specially named repository path that includes Unicode characters. That combination causes Ruby to recompile the expression after GitLab’s own size validation has already passed, triggering an integer overflow that redirects the regex engine’s execution into a command-execution function.

Critically, the access bar for triggering this chain is lower than many critical vulnerabilities require. The attacker only needs permission to push a branch to a public project on the target instance — something that’s often trivially available on self-hosted or loosely configured GitLab deployments. From there, a follow-up request to the CI lint endpoint requires no authentication token, session cookie, or other authorization header at all.

What Successful Exploitation Looks LikeBurgess notes that commands triggered through the bug execute under GitLab’s internal git service account rather than as a system administrator — a meaningful limitation, but still enough to give an attacker a foothold for lateral movement, source code access, or further privilege escalation depending on the server’s configuration.

Detecting a successful attempt isn’t entirely straightforward either. The test request typically ends in an HTTP 502 error as the worker process crashes after the injected command runs, which on its own doesn’t confirm exploitation occurred. In Burgess’s testing, confirmation came through command output returned via an external callback channel. The exploit also demands significant server memory to work reliably; under low-memory conditions or concurrent requests, it’s more likely to cause a denial-of-service crash than clean code execution. The published PoC is calibrated specifically against the official GitLab CE 19.4.1 Linux amd64 image, so results may vary against other builds.

Which Versions Need PatchingGitLab’s advisory lists the affected ranges precisely:

19.2 before 19.2.8

19.3 before 19.3.5

19.4 before 19.4.2

Fixed versions were released on October 12, 2026. GitLab.com is already running a patched build, and GitLab Dedicated customers don’t need to take any action. The risk falls squarely on self-managed instances that haven’t yet applied the update.

Recommended ActionsWith exploit code already public, GitLab administrators running self-managed instances should move quickly:

Confirm your installed version against the affected ranges above and upgrade immediately to 19.2.8, 19.3.5, 19.4.2, or later.

Review recent CI/CD configuration changes across projects, paying particular attention to new or modified regular expressions and unusual branch names containing non-standard characters.

Audit who has permission to push branches to public projects, since that’s the only access requirement for the first stage of the attack.

Watch for unexplained 502 errors tied to CI lint requests, which could indicate exploitation attempts in progress.

There’s no confirmed evidence yet that CVE-2026-97229 is being exploited in real-world intrusions. But a maximum CVSS score combined with public, working exploit code and a low bar for initial access is exactly the combination that tends to draw opportunistic attackers quickly. For any organization running GitLab’s CI/CD pipelines in production, this patch shouldn’t wait for the next scheduled maintenance window.

#gitlab #rce #vulnerability

0

Caricamento...

0
0

Caricamento...

OpenAI Account Takeover via libheif Heap Overflow and SSO Flaw: How Hacktron Chained a Malicious HEIF Image to Internal Repo Access — Analysis & Lesso

Altro...

OpenAI Account Takeover via libheif Heap Overflow and SSO Flaw: How Hacktron Chained a Malicious HEIF Image to Internal Repo Access — Analysis & Lessons#cybersecurity
OpenAI Account Takeover via libheif Heap Overflow and SSO Flaw: How Hacktron Chained a Malicious HEIF Image to Internal Repo Access — Analysis & Lessons

#cybersecurity #software #infosec #coding #community #engineering #development #inclusive #vulnerability #bugbounty

0

Caricamento...

0
1

Caricamento...

CVE-2026-105192: LMCache (vLLM KV-Cache) Unauthenticated Remote Code Execution Vulnerability — Analysis & Mitigation#cybersecurity

Altro...

CVE-2026-105192: LMCache (vLLM KV-Cache) Unauthenticated Remote Code Execution Vulnerability — Analysis & Mitigation#cybersecurity
CVE-2026-105192: LMCache (vLLM KV-Cache) Unauthenticated Remote Code Execution Vulnerability — Analysis & Mitigation

#cybersecurity #software #infosec #coding #cve #community #engineering #development #inclusive #vulnerability

0

Caricamento...

0
1

Caricamento...

CVE-2026-96951: Microsoft Windows Shell Mark-of-the-Web Bypass Vulnerability — Analysis & Mitigation#cybersecurity

Altro...

CVE-2026-96951: Microsoft Windows Shell Mark-of-the-Web Bypass Vulnerability — Analysis & Mitigation#cybersecurity
CVE-2026-96951: Microsoft Windows Shell Mark-of-the-Web Bypass Vulnerability — Analysis & Mitigation

#cybersecurity #software #infosec #coding #cve #community #engineering #development #inclusive #vulnerability

0

Caricamento...

0
1

Caricamento...

CVE-2026-104935: Ivanti Connect Secure Authentication Bypass Vulnerability — Analysis & Mitigation#cybersecurity

Altro...

CVE-2026-104935: Ivanti Connect Secure Authentication Bypass Vulnerability — Analysis & Mitigation#cybersecurity
CVE-2026-104935: Ivanti Connect Secure Authentication Bypass Vulnerability — Analysis & Mitigation

#cybersecurity #software #infosec #coding #cve #community #engineering #development #inclusive #vulnerability

0

Caricamento...

0
1

Caricamento...

Microsoft’s October Patch Tuesday Fixes Over 1,100 Flaws and Debuts a New Way to Triage Them

Read Time:3 Minute, 29 Second

A Record-Sized Update BatchMicrosoft’s October 2026 Patch Tuesday release addressed 1,118

Altro...

Read Time:3 Minute, 29 Second

A Record-Sized Update BatchMicrosoft’s October 2026 Patch Tuesday release addressed 1,118 vulnerabilities, one of the company’s largest monthly security updates to date, continuing a trend of expanding update volumes seen in its September and August 2026 releases. For enterprise security teams already stretched thin by routine patching cycles, a release of this size can turn a monthly chore into a weeks-long undertaking.

Introducing “Risk Cohorts”To help ease that burden, Microsoft is piloting a new organizational concept it calls risk cohorts: groupings of vulnerabilities that share the same underlying exposure questions, rather than groupings based purely on severity score or affected product. The idea is that many CVEs, even unrelated ones, ultimately require security teams to answer a similar handful of questions, such as whether untrusted files can reach a particular Windows component, or whether a lower-privileged user can execute code locally.

By grouping flaws this way, Microsoft says analysts can answer a shared question once for an entire cohort rather than repeating nearly identical investigative work across dozens or hundreds of individual CVEs. Crucially, Microsoft stresses that cohorts are explicitly exploratory and not official vulnerability categories: they do not replace the Security Update Guide, do not establish urgency on their own, and do not remove the need to check exploitation status, required user interaction, or available attack paths for each flaw individually.

Where the Vulnerabilities Are ConcentratedFifteen cohorts were defined for October, and Windows- and Office-related issues dominate the list by volume. The largest, covering Windows filesystems, storage, and backup features, accounts for 136 CVEs and centers on which features handle untrusted local input, disk images, or media. Office document and mail clients follow closely with 135 CVEs, where the key question is which applications are installed, what untrusted content reaches them, and how updates are deployed. Windows graphics, media, and content-processing components make up a further 131 flaws.

Windows core and kernel privilege boundaries: 109 CVEs

Installation, deployment, and update services: 103 CVEs

Device, peripheral, and print surfaces: 76 CVEs

DNS, DHCP, and network-address infrastructure: 76 CVEs

Remote-access and RDP deployments: 65 CVEs

Authentication, credentials, and cryptography: 61 CVEs

Windows networking and protocol stack: 53 CVEs

Virtualization, container, and isolation hosts: 45 CVEs

Network file and resource sharing: 39 CVEs

Endpoint management and security controls: 38 CVEs

Windows web, RPC, and messaging services: 27 CVEs

Directory, federation, and certificate infrastructure: 24 CVEs

What Security Teams Should Actually DoFor each cohort, Microsoft provides a shared assessment question meant to streamline triage. For the remote-access and RDP group, for instance, the relevant question is which hosts function as clients, gateways, or servers, and which identities and access paths are permitted to reach them. For virtualization and isolation, the question is where untrusted guest or workload code is allowed to execute.

That framing can speed up the first pass of a large release, but Microsoft is careful to note its limits. Two vulnerabilities sitting in the same cohort can still carry very different severity ratings, exploitation requirements, or exposure levels, one might require authenticated local access while another is remotely exploitable with no user interaction at all. The company’s existing Security Update Guide remains the authoritative source for the specifics of any individual CVE, with risk cohorts meant to sit alongside that process rather than replace it.

A Response to Patching FatigueThe experiment reflects a broader industry challenge: as vendors ship larger and more frequent security updates, the manual effort required to triage, prioritize, and deploy them is growing faster than most security teams’ capacity to keep up. Grouping related questions is a modest but practical step toward making huge releases more manageable, even if it does not reduce the underlying number of fixes that eventually need to be applied.

What To Do NowOrganizations should treat October’s release with the same urgency as any other Patch Tuesday: prioritize systems exposed to the internet or accessible to lower-privileged users, apply updates addressing actively exploited or publicly disclosed flaws first, and use Microsoft’s cohort mapping as a starting point for triage rather than a final verdict on priority. Teams unfamiliar with the new framework should review Microsoft’s published cohort documentation alongside the standard Security Update Guide before adjusting existing patch management workflows.

#microsoft #windows #patchtuesday #vulnerability

0

Caricamento...

0
0

Caricamento...

Thousands of Exposed NVIDIA GPU Monitors Leave AI Data Centers Open to Crash Attacks

Read Time:3 Minute, 45 Second

A high-severity vulnerability in one of the most widely used monitoring tools for NVIDIA

Altro...

Read Time:3 Minute, 45 Second

A high-severity vulnerability in one of the most widely used monitoring tools for NVIDIA GPU fleets has left thousands of AI data centers and Kubernetes clusters open to disruption. Security researchers tracking internet-facing infrastructure found more than 2,000 exposed instances of NVIDIA’s DCGM Exporter, a discovery that highlights how a single overlooked debugging feature can put critical AI infrastructure at risk.

A Monitoring Tool Built for Insight, Not ExposureDCGM Exporter is the go-to utility for teams running large GPU fleets. It pulls live metrics—temperature, memory usage, power draw, error counts, and workload activity—straight from NVIDIA hardware and republishes them in a format that Prometheus and similar monitoring stacks can scrape over plain HTTP. Each GPU is tagged with a unique identifier, which makes the exporter extremely useful for operators but equally useful to an outsider who stumbles onto an unsecured endpoint.

If that endpoint sits open on the public internet, anyone can map an organization’s hardware inventory, observe usage patterns, and infer details about the workloads running behind it. That alone is a privacy and operational security concern, but the newly disclosed flaw goes a step further by letting an attacker take the monitoring down entirely.

The Flaw: Unlimited Access to Debug RoutesTracked as CVE-2026-47483 and rated 8.2 out of 10 in severity, the bug lives in the exporter’s Go-language profiling routes, exposed under the /debug/pprof path. These routes exist so developers can inspect CPU and memory behavior during troubleshooting, but in vulnerable builds they shipped without authentication or rate limiting. NVIDIA classified the issue as CWE-770, a textbook case of resource allocation without proper limits.

That gap means an attacker needs no credentials at all. By firing off a burst of profiling requests, they can force the exporter into runaway memory and CPU consumption until it slows to a crawl, stops responding, or crashes outright.

What Happens When the Exporter Goes DownLosing DCGM Exporter is not a cosmetic problem. A crash wipes out the live view into GPU health, meaning operators lose visibility into temperature spikes, memory errors, and failing hardware exactly when they would need it most. On hosts without strict container or process limits, the same resource pressure can bleed into neighboring AI workloads running on the same machine, slowing down training or inference jobs that have nothing to do with the attack itself.

Researchers at Lava Security, who published the technical analysis behind the discovery, say they only tested the resource-exhaustion behavior in a controlled lab and did not interfere with any of the live systems they located during their internet-wide scan.

A Surprisingly Large Attack SurfaceUsing Shodan, the researchers identified upward of 2,000 exposed DCGM Exporter hosts spanning nearly 300 separate organizations, collectively revealing more than 12,000 unique GPU identifiers reachable over unauthenticated HTTP. Given how central GPU capacity has become to AI development, that is a notable number of organizations effectively broadcasting details about their compute infrastructure—and leaving a crash switch open to the public—without realizing it.

Recommended MitigationsNVIDIA has shipped a patch, but rolling it out across every node in a fleet is the part organizations tend to miss. Security teams should treat this as a two-part fix: patch the software, and lock down network exposure so a similar debug-route mistake can’t be exploited again. Specific steps include:

Bind DCGM Exporter, Prometheus, and Node Exporter services to a private or loopback interface rather than a public one.

Restrict access to approved monitoring systems only, enforced through firewall rules, cloud security groups, or Kubernetes NetworkPolicies.

Audit internet-facing port 9400 and block public access to the /debug/pprof path specifically.

Watch for sudden memory growth, CPU spikes, failed Prometheus scrapes, or unexplained gaps in GPU telemetry—all of which can indicate probing or active exploitation.

Verify that the patched container image is actually deployed to every GPU node, not just updated in a registry or a configuration file that hasn’t been rolled out yet.

The underlying lesson extends well beyond this one exporter: any internal debugging or profiling interface that gets bundled into a production container image is a liability the moment that container faces the public internet. As AI infrastructure scales out across cloud and on-premises GPU clusters, auditing exactly what each monitoring sidecar exposes—and to whom—deserves the same scrutiny as the workloads it’s meant to observe.

#nvidia #aiinfrastructuresecurity #vulnerability

0

Caricamento...

0
0

Caricamento...