CVE-2026-125350: Chrome XML use-after-free, now actively exploited and on CISA's KEV with a November 24 deadline

A use-after-free in Chrome's XML handling is under active exploitation per CISA. Update every Chromium-based browser to Chrome 157.0.8103.50 or later before the November 24 federal deadline.
CoeLab Real-World Impact: 87/100 — Urgent. Editorial score v1.0. Evidence checked 2026-11-12T16:10:47.508962+00:00. CVSS and EPSS are separate measures, not this editorial score.
Rating breakdown: exploitation: 30, reachability: 25, consequences: 22, operational relevance: 10.
Who is affected
CVE-2026-125350 is a use-after-free vulnerability in the XML component of Google Chrome. Google's own Chrome security team is the CVE numbering authority for this record, and the affected-version data is unambiguous: every Chrome release before 157.0.8103.50 is affected. Chrome 157.0.8103.50 is the fixed version.
The practical blast radius is larger than Chrome alone. The CISA Known Exploited Vulnerabilities entry, which mirrors Google's advisory, notes that this vulnerability affects "multiple web browsers that utilize Chromium, including, but not limited to, Google Chrome, Microsoft Edge, and Opera." Chromium-derived browsers follow Chromium's release trains, so organizations that standardize on Edge or other Chromium-based browsers should treat the same fix urgency as if the bug carried their browser's own name. Firefox is built on a different engine and is not implicated by this record.
Desktop distributions are the class named in the advisory ("Stable Channel Update for Desktop"). Because the defect lives in the shared XML code of the browser engine, it is not confined to a specific platform: Windows, macOS, and Linux desktop builds of the affected versions are all in scope.
There is no meaningful distinction between "work" and "home" exposure here. Any user who browses the web with an unpatched Chromium-based browser is a candidate target, because the trigger is a crafted web page.
Real-world impact and exploitation evidence
If exploitation succeeds, the attacker gains the ability to execute arbitrary code inside the browser's sandbox. The CVE description is explicit about the boundary: code executes "inside the sandbox via a crafted HTML page." That is an important qualifier. The victim's machine is not immediately owned; the attacker first gets a foothold in the renderer sandbox, one of the most hardened layers of the browser.
Two things nonetheless make this serious. First, an attacker with sandboxed code execution has broken the isolation between untrusted web content and the rest of the browser's renderer process — a prerequisite step for the class of attack chains that pair a browser bug with a sandbox-escape bug to reach the operating system. Second, and most decisively, CISA's SSVC assessment for this record — authored by the CISA Coordinator and attached to the CVE record on 2026-11-11 — marks Exploitation as "active" with "Technical Impact: total." Google awarded the bug its top Chromium security severity of High and patched it within a day of the affected release.
The exploitation evidence timeline is what turns a routine patch into urgent news. CVE data was reserved on 2026-11-09, the Program record was published on 2026-11-10, and CISA added the vulnerability to the KEV catalog the same day, 2026-11-10. KEV listing is an evidence-based U.S. government determination that the vulnerability has been exploited in the wild — it is not a prediction or a heuristic. From our review of the record and the KEV entry, the concrete exploitation window at publication of the CVE record is evidence that exploitation occurred by that date; we cannot determine from primary sources exactly when in-the-wild exploitation first began or its scope.
CISA has set a federal remediation deadline of 2026-11-24 under BOD 26-04, giving agencies roughly two weeks from the KEV addition. For everyone else, the deadline should be treated as "today," because the window between an actively exploited, publicly documented bug and mass-available patches closing at 157.0.8103.50 is exactly when opportunistic exploitation peaks.
FIRST EPSS currently estimates a roughly 1% probability of exploitation attempts in the next 30 days (EPSS 0.0096, 62nd percentile, dated 2026-11-12). That number understates observed reality here — EPSS is a statistical model, and KEV listing plus CISA's "active" SSVC determination are direct evidence that beat a model prediction. Never read a low EPSS score as "not exploited" when the KEV catalog says otherwise.
Attack prerequisites
The attack vector is the network: an attacker must get a victim to load attacker-controlled HTML content. The CVSS 3.1 vector published in the CVE record — CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H with a base score of 8.8 (High) — spells out the prerequisites precisely: no privileges are required (PR:N), the complexity of the attack is low (AC:L), but user interaction is required (UI:R). The scope of the impact is the vulnerable component itself (Scope: Unchanged), with high impact to confidentiality, integrity, and availability once code runs.
In plain terms, the attacker needs to lure or redirect a victim to a malicious web page — through a phishing link, compromised ad content, a poisoned search result, or an injected element on an otherwise legitimate site. No authentication, no VPN access, no internal foothold is needed first. That combination — remote, unauthenticated, low complexity — is why browser bugs dominate exploitation statistics: the entire delivery mechanism is "somebody browsed the wrong page."
What the prerequisites do not include is any information leak or chain partner documented as required. The record describes a single self-contained bug. That said, the sandbox confinement described above means that an attacker seeking full system compromise would in practice want an additional sandbox-escape bug; nothing in the primary sources we reviewed documents such a companion bug for this CVE, and we decline to speculate about attack chains not described by Google or CISA.
Actions and fixes
The fix is deterministic and already available: update Chrome to 157.0.8103.50 or later. Google shipped this fix in the Stable Channel update for desktop referenced in the CVE record, and all Chromium-based browsers inheriting the fix should be updated to their own vendor's release carrying Chromium 157.0.8103.50-equivalent code. In Edge, Opera, Brave, and similar browsers, check the browser's built-in version display and allow the browser to auto-update, then restart the browser fully — a use-after-free fix in the renderer only takes effect after every renderer process has restarted, which means closing all browser windows.
Until every managed endpoint is patched, the practical compensating controls are limited but real: enforce browser auto-update policies centrally; block or sandbox browsing on high-risk endpoints where feasible; and treat any endpoint that visited untrusted sites with an unpatched browser as a candidate for forensic review, per the guidance CISA's KEV entry references under its BOD 26-04 forensics triage requirements. CISA's required action for federal agencies directs mitigation per vendor instructions within the deadline — the vendor instruction here is simply "update to the fixed release."
There is no configuration change or registry setting documented in the primary sources that safely disables the affected XML processing path, so we do not recommend attempting one. The only sound mitigation is the update itself.
One operational note: because KEV entries drive automated compliance scanning in many organizations, expect vulnerability scanners to flag anything below the fixed version immediately. Budget for the churn — the correct response is to close the gap, not the alert.
Uncertainty and what could change
Several facts cannot be determined from the primary sources available at the time of writing ( November 12, 2026): the date of first in-the-wild exploitation, who is exploiting the bug and against whom, how many organizations have observed it, and whether any attack chain pairs it with a sandbox escape for full system compromise. Google's issue tracker entry for this bug (issue 582442907) is access-restricted, so no additional technical detail was independently verifiable. If CISA or Google later publishes attribution or campaign details, the urgency calculus could shift again — most likely upward.
The affected-version boundary is exactly as Google states it: below 157.0.8103.50 is affected, this version is the fix. We did not independently test or verify behavior on any specific browser build or platform, and readers should rely on their browser vendors' own updates for derivatives. Extensions, mobile WebView, and embedded Chromium frameworks are not addressed in this record and we make no claim about them.
EPSS will move as attackers and defenders react; a rising score in the coming days would be consistent with the KEV determination, not new evidence. The KEV due date (2026-11-24) is a compliance deadline, not a prediction that exploitation will stop or start on any date.
Finally, we assess the CoeLab Real-World Impact score, not CVSS or EPSS, as our editorial severity signal. This CVE scored 87 of 100 ("Urgent") under v1.0. Its components: full 30/30 exploitation points because the KEV listing with an active-exploitation determination is direct evidence of in-the-wild use; 25/25 reachability from the network vector with no privileges and low complexity; 22/25 consequences from the published CVSS 8.8; and only 10/20 operational relevance from the generic "common platform" category match. The score's limitations matter here: it has no measured installed-base data on our side, treats any Chromium-based browser as a common platform rather than measuring actual deployments we can observe, and the exploitation component is binary on KEV membership. It is not a CVSS substitute and not an EPSS substitute, and it says nothing about any specific reader's environment. A score of 87 is strong support for publishing — it is not, by itself, proof of the scale of exploitation, which is exactly what we could not establish from primary sources.
Sources and evidence date
This analysis is based on the CVE Program record for CVE-2026-125350 (CNA: Chrome, published 2026-11-10, last updated 2026-11-11), the CISA Known Exploited Vulnerabilities catalog entry dated 2026-11-10 with a remediation due date of 2026-11-24, and Google's Chrome Releases Stable Channel Update for Desktop referenced by both. We reviewed the CISA Coordinator's SSVC assessment attached to the CVE record and the FIRST EPSS data point dated 2026-11-12.
Availability note: the referenced Chromium issue tracker entry requires sign-in, and Google's Chrome Releases blog entry was machine-fetched without full article content, so we relied on the CVE Program record as the authoritative statement of affected and fixed versions. All version claims in this article come from that record. Evidence was gathered on 2026-11-12 at approximately 19:45 UTC; anything published by vendors or CISA after that time is not reflected here.
This article is published by The Coe Lab. Editorial contact: https://thecoelab.com/contact
Primary-source links
https://chromereleases.googleblog.com/2026/11/stable-channel-update-for-desktop_018223192.html
https://www.cve.org/CVERecord?id=CVE-2026-125350
https://www.cisa.gov/known-exploited-vulnerabilities-catalog?field_cve=CVE-2026-125350
https://issues.chromium.org/issues/582442907
AI-assisted research and drafting, with a separate model editorial review. CoeLab ratings are editorial judgments, not a substitute for checking your own versions and vendor advisories. Methodology: https://thecoelab.com/blog/security/methodology
Related Posts
Your USB-C Dock Is Part of Your Network. Test It That Way.
A trending Ethernet investigation shows why docks belong in your network diagram. Here is how to isolate bottlenecks before replacing routers or adapters.
CVE-2026-129448: Apache Roller XML-RPC deserialization gets CISA KEV deadline — unauthenticated remote code execution in versions before 6.1.5
Apache Roller before 6.1.5 runs unauthenticated XML-RPC deserialization before authentication; it is now on CISA's exploited list with a patch deadline of November 23, 2026.
Valkey’s Sorted Set Upgrade Is a Capacity Planning Opportunity
Valkey’s sorted set redesign promises lower index overhead. Here is how to test the real memory benefit, protect latency, and turn savings into usable capacity.