← Back to Feed CACHED · 2026-05-17 09:42:19 · cache_key CVE-2025-29912
CVE-2026-11085 · CWE-472 · Disclosed 2026-06-04

Integer overflow in GPU in Google Chrome on Android prior to 149

ASSESSED — NOISGATE V0.5
Vendor
Reassessed
Verdict:
01 · The Real Story

This is a second lock behind a first broken door, dangerous in a full break-in but rarely the thing that gets opened first

The authoritative record matching your description appears to be CVE-2026-11256, not CVE-2026-11085. It is an integer overflow in the GPU component of Google Chrome on Android affecting versions before 149.0.7827.53. The key detail is the one that changes everything operationally: the attacker must have already compromised the renderer process and then use a crafted HTML page to try to turn that foothold into a sandbox escape.

Your supplied HIGH 8.8 framing overstates the real enterprise urgency. The vendor language says Chromium security severity: Low, and while CISA-ADP later attached a high CVSS-style score, that numeric model still treats the bug too much like a first-hop remote compromise. In the real world this is a post-renderer-compromise chain element on Android only, with no KEV listing, no public exploitation evidence, and extremely low EPSS, so the right move is to downgrade it to a routine browser-update priority rather than a fleet-wide fire drill.

"This is a chainable sandbox-escape component, not a stand-alone internet-to-root emergency."
02 · The Attack Path

4 steps from start to impact.

STEP 01

Land a renderer foothold with a separate browser bug

A real attacker does not start with this CVE; they start by gaining code execution or equivalent control inside Chrome's renderer through some other memory-corruption or logic bug. This CVE only becomes relevant after that first stage has already succeeded. *Weaponized tool:* a separate renderer exploit chain or 1-day, not a public PoC for this CVE itself.
Conditions required:
  • Victim uses vulnerable Chrome on Android prior to 149.0.7827.53
  • Victim is lured to attacker-controlled content
  • Attacker has a separate renderer-compromise primitive
Where this breaks in practice:
  • Requires a second vulnerability or pre-existing compromise before this CVE matters
  • Modern Safe Browsing, site isolation, and renderer hardening raise the bar for the prerequisite stage
  • No public PoC or exploit chain was found for this specific issue
Detection/coverage: Standalone vuln scanners usually see only version exposure. They do not prove exploitability because the decisive prerequisite is a separate renderer compromise.
STEP 02

Trigger the GPU integer overflow from crafted web content

Once the renderer is compromised, the attacker drives the GPU path with crafted HTML intended to exercise the integer-overflow condition. The overflow is useful because it may create an out-of-bounds memory condition in a privileged browser component. *Weaponized tool:* custom HTML/JS exploit stage delivered from the attacker page.
Conditions required:
  • Renderer process is already controlled or materially compromised
  • GPU code path is reachable on the target device and build
  • The specific vulnerable code path is still present in the installed version
Where this breaks in practice:
  • Exploit reliability on Android GPU paths is often device/build sensitive
  • Chrome bug details are still restricted, limiting commodity weaponization
  • Attack complexity is reflected as high by ADP
Detection/coverage: Network controls will not reliably flag this; telemetry is more likely to come from browser crash spikes, exploit-protection hits, or vendor threat intel rather than IDS signatures.
STEP 03

Turn browser-component memory corruption into sandbox escape

The attacker then attempts to convert memory corruption in the GPU path into execution outside the compromised renderer sandbox. That is the actual security value of this CVE: escaping a constrained renderer into a more privileged browser context. *Weaponized tool:* a custom sandbox-escape payload; no public exploit kit reference found.
Conditions required:
  • Precise memory corruption control sufficient for exploitation, not just crash
  • Target device/browser mitigations can be bypassed
  • Browser sandbox boundaries remain the next reachable privilege tier
Where this breaks in practice:
  • Many memory bugs crash far more often than they cleanly escape sandbox
  • Exploit must survive modern mitigations and Android variance
  • Impact is significant only if the attacker already solved stage one
Detection/coverage: EDR coverage on Android is limited in many enterprises. Detection is strongest through managed mobile telemetry, abnormal browser crashes, and vendor/browser threat reporting.
STEP 04

Use the escaped context for follow-on abuse

After a successful escape, the attacker can pursue data theft, persistence within browser context, or further privilege escalation using additional local primitives. This CVE alone does not guarantee full device compromise. *Weaponized tool:* follow-on post-exploitation chain, which would vary by campaign.
Conditions required:
  • Sandbox escape succeeded
  • There is valuable enterprise data in browser/session context
  • Additional controls such as app isolation or MDM policies do not blunt impact
Where this breaks in practice:
  • Blast radius is limited to Android Chrome users, not your server estate
  • Many enterprises keep mobile browsers low-privilege and containerized
  • No evidence this CVE is being operationalized in the wild today
Detection/coverage: Look for anomalous session hijack indicators, managed mobile-browser telemetry, and repeatable crash/upgrade lag cohorts rather than perimeter scanning.
03 · Intelligence Metadata

The supporting signals.

Record sanity checkThe description you supplied matches CVE-2026-11256, not CVE-2026-11085. I assessed the authoritative record matching the description.
In-the-wild statusNo current exploitation evidence found and not KEV-listed in the CISA catalog.
PoC availabilityNo public PoC found. The Chromium bug reference is still restricted: issues.chromium.org/issues/498856565. That meaningfully slows commodity weaponization.
EPSS0.00068 probability, 0.211 percentile on 2026-06-05 via CVE.report's FIRST-fed aggregation. That's effectively background noise, not an exploitation magnet.
KEV statusNo. As of the current CISA KEV catalog, this issue is absent.
CVSS reality checkADP shows 8.3 / CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H on NVD, but the practical hidden prerequisite is renderer already compromised. That is major downward pressure in real fleets.
Affected versionsGoogle Chrome on Android before 149.0.7827.53 per NVD.
Fixed versionUpdate to 149.0.7827.53 or later. Chrome Android stable updates are distributed via Google Play; Chrome release notes show Android 149 rollout on May 28, 2026.
Vendor-vs-ADP severityVendor text says 'Chromium security severity: Low' on NVD, while CISA-ADP later attached a HIGH 8.3 vector. For patch triage, the vendor's *chain-element* framing is closer to reality.
Scanning / exposure dataNot Shodan/Censys/GreyNoise-friendly. This is a client-side mobile-browser bug, not an internet-listenable service. Exposure measurement is an asset inventory / app-version problem, not an external-attack-surface problem.
04 · The Call

noisgate verdict.

Final Verdict
DOWNGRADED to MEDIUM (5.6/10)

The decisive factor is the prerequisite chain: this bug is useful only after the attacker has already compromised the Chrome renderer. That makes it a post-initial-compromise sandbox escape on Android, not a broad unauthenticated fleet-wide entry point, which is why it lands in MEDIUM despite scary memory-corruption language.

HIGH Downward pressure from the 'renderer already compromised' prerequisite
MEDIUM Assessment that no public PoC or active exploitation is currently driving urgency
MEDIUM Mapping the user's CVE ID to the authoritative record matching the supplied description

Why this verdict

  • Requires prior compromise: an attacker must already control or materially compromise the renderer process before this CVE does anything useful.
  • Android-only population: this is not your Windows/Mac/Linux browser fleet; it is a narrower mobile subset with smaller enterprise blast radius.
  • No live-fire intel: no KEV listing, no public exploitation evidence, and a near-floor EPSS score argue against emergency treatment.
  • Version exposure is common, exploitability is not: lots of devices may be behind on Chrome, but very few real attackers can chain a renderer bug plus a reliable Android GPU sandbox escape.

Why not higher?

This is not a stand-alone drive-by RCE from the open internet. The attack path assumes a prior renderer break, plus reliable exploitation of a GPU memory bug on Android, plus successful sandbox escape. Each prerequisite sharply narrows the number of attackers and targets that can operationalize it.

Why not lower?

It is still a memory-corruption bug in a massively deployed browser component, and its intended end state is a sandbox escape. If you operate a large managed Android fleet, especially with sensitive browser-based workflows, this is worth fixing through normal browser-update discipline rather than treating it as pure backlog trivia.

05 · Compensating Control

What to do — in priority order.

  1. Enforce Chrome auto-update through EMM/Managed Google Play — Make the managed Play policy for com.android.chrome mandatory and verify update deferrals are not pinning devices below 149.0.7827.53. For a MEDIUM verdict there is no mitigation SLA — go straight to the 365-day remediation window, but this control is low-cost and should be implemented during the next routine mobile policy review.
  2. Inventory Android Chrome versions daily — Treat this as an asset-version hygiene problem, not a perimeter exposure problem. Pull app version telemetry from your MDM/UEM and alert on devices below 149.0.7827.53; for MEDIUM, there is no mitigation SLA, so use this to drive orderly remediation inside the 365-day window.
  3. Reduce unmanaged browsing paths — Where possible, require corporate browsing through managed Chrome and block alternate unmanaged browsers for sensitive workflows. That does not remove the bug, but it shrinks the population you must trust and gives you version visibility; again, no mitigation SLA applies here, so fold it into normal mobile hardening work.
  4. Monitor crash and exploit telemetry on mobile — Watch for clusters of Chrome crashes, browser instability, or vendor/mobile threat-defense alerts that could indicate exploit development against GPU paths. There is no mitigation SLA for this MEDIUM verdict, but telemetry is your early-warning system if the threat picture changes and urgency needs to be re-rated.
What doesn't work
  • A network IDS signature will not save you here; the exploit would ride normal web content and the decisive stage is client-side memory corruption.
  • External attack-surface scanners like Shodan/Censys are basically irrelevant; this is not a remotely fingerprintable listener.
  • WAF rules do not meaningfully help because the victim is the browser engine, not your web server.
06 · Verification

Crowdsourced verification payload.

Run this on an auditor workstation with Android platform-tools (adb) installed, targeting a managed Android device with USB debugging or enterprise ADB access enabled. Invoke it as ./check_chrome_android.sh <device-serial> for a specific device or ./check_chrome_android.sh when only one device is attached; no root is required on the phone, but you need permission to query package metadata over ADB.

noisgate-verify.sh
BASHREAD-ONLYSAFE
#!/usr/bin/env bash
# check_chrome_android.sh
# Verifies whether Google Chrome on an attached Android device is below 149.0.7827.53.
# Output: VULNERABLE / PATCHED / UNKNOWN
# Exit codes: 0=PATCHED, 1=VULNERABLE, 2=UNKNOWN

set -u

FIXED_VERSION="149.0.7827.53"
PACKAGE="com.android.chrome"
SERIAL="${1:-}"

adb_cmd() {
  if [ -n "$SERIAL" ]; then
    adb -s "$SERIAL" "$@"
  else
    adb "$@"
  fi
}

version_lt() {
  # returns 0 if $1 < $2
  local IFS=.
  local i
  local -a a=($1) b=($2)
  local len=${#a[@]}
  if [ ${#b[@]} -gt $len ]; then len=${#b[@]}; fi
  for ((i=0; i<len; i++)); do
    local ai=${a[i]:-0}
    local bi=${b[i]:-0}
    if ((10#$ai < 10#$bi)); then
      return 0
    elif ((10#$ai > 10#$bi)); then
      return 1
    fi
  done
  return 1
}

# Basic device check
if ! adb_cmd get-state >/dev/null 2>&1; then
  echo "UNKNOWN - no accessible Android device via adb"
  exit 2
fi

# Try several methods because OEM builds differ
VERSION="$(adb_cmd shell dumpsys package "$PACKAGE" 2>/dev/null | tr -d '\r' | sed -n 's/.*versionName=//p' | head -n1)"
if [ -z "$VERSION" ]; then
  VERSION="$(adb_cmd shell cmd package list packages -f "$PACKAGE" 2>/dev/null | tr -d '\r' | sed -n 's/.*versionName=//p' | head -n1)"
fi
if [ -z "$VERSION" ]; then
  VERSION="$(adb_cmd shell pm dump "$PACKAGE" 2>/dev/null | tr -d '\r' | sed -n 's/.*versionName=//p' | head -n1)"
fi

if [ -z "$VERSION" ]; then
  # Final fallback: try package manager path presence to distinguish absent vs unreadable
  if adb_cmd shell pm path "$PACKAGE" >/dev/null 2>&1; then
    echo "UNKNOWN - Chrome installed but version could not be read"
  else
    echo "UNKNOWN - Chrome is not installed or package metadata is inaccessible"
  fi
  exit 2
fi

if version_lt "$VERSION" "$FIXED_VERSION"; then
  echo "VULNERABLE - $PACKAGE version $VERSION is below fixed $FIXED_VERSION"
  exit 1
else
  echo "PATCHED - $PACKAGE version $VERSION is at or above fixed $FIXED_VERSION"
  exit 0
fi
07 · Bottom Line

If you remember one thing.

TL;DR
Monday morning: do not treat this like a fleet emergency. First, correct the record in your tracker to the authoritative issue matching the supplied description (CVE-2026-11256), then pull an MDM/UEM report for Android devices running Chrome below 149.0.7827.53 and make sure managed Play auto-update policy is actually landing. Because this lands at MEDIUM and there is no KEV or active exploitation evidence, there is noisgate mitigation SLA — go straight to the 365-day remediation window; use your noisgate remediation SLA to get vulnerable Android Chrome installs updated within 365 days, but in practice fold it into your next normal mobile browser update cycle rather than burning an emergency patch window.

Sources

  1. NVD detail for the authoritative matching record
  2. Chrome Early Stable desktop update 149.0.7827.53/.54
  3. Chrome for Android update 149.0.7827.48 rollout
  4. Chromium issue tracker reference for this bug
  5. CISA Known Exploited Vulnerabilities catalog
  6. FIRST EPSS project overview
  7. CVE.report aggregation including EPSS snapshot
Peer Review

What defenders are saying.

Submit a review attribution: handle + country only
0 flags selected · stored anonymously
Validation Results

Crowdsourced verification outputs.

Results submitted by users who ran the verification payload against their environment.