← Back to Feed CACHED · 2026-05-17 09:42:19 · cache_key CVE-2025-29912
CVE-2026-10900 · CWE-416 · Disclosed 2026-06-04

Use after free in Passwords in Google Chrome on Mac prior to 149

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

This is less a front-door smash and more a booby-trapped drawer that only opens after a fussy sequence of clicks

Per the supplied intel, CVE-2026-10900 is a use-after-free in Chrome's Passwords component on macOS affecting Google Chrome on Mac before 149.0.7827.53, disclosed on 2026-06-04. The described path is not a normal drive-by page load; the attacker has to get a user onto malicious content and then induce specific UI gestures, which strongly suggests the vulnerable code lives behind password-management or browser-UI flows rather than routine rendering alone.

Google's HIGH / 7.5 baseline is defensible if you score the bug in a vacuum as browser memory corruption with full CIA impact. In enterprise reality, that score overshoots: macOS-only, user interaction required, high attack complexity, no KEV, tiny EPSS, and no public exploitation evidence all compress the reachable population. This still matters because browser memory corruption is never backlog trash, but it is not the kind of fleet-wide emergency that a broad drive-by Chrome RCE would be.

"Dangerous bug class, but the real-world path is too narrow for a top-tier patch panic."
02 · The Attack Path

4 steps from start to impact.

STEP 01

Land the user on attacker-controlled content

The attacker first needs the victim to browse to malicious web content or a lure page that can steer them toward the Passwords-related UI path. This is an internet-origin attack surface, but it is not enough by itself; the page must also get the user to do more than simply load it.
Conditions required:
  • Victim uses Chrome on macOS
  • Victim is running a version earlier than 149.0.7827.53
  • Attacker can deliver a convincing web lure
Where this breaks in practice:
  • Chrome auto-update shrinks dwell time on unmanaged consumer-style installs
  • Managed enterprise Macs often restart Chrome regularly via MDM or user workflows
  • This does not appear to trigger from passive page render alone
Detection/coverage: Version scanners and EDR inventory can identify vulnerable installs well; network detection is weak because the initial step looks like ordinary web browsing.
STEP 02

Drive the victim into the narrow UI path

The supplied title says the attacker must convince the user to engage in specific UI gestures. That is a major real-world constraint: this likely means opening a password-related surface, clicking through a specific sequence, or interacting with browser chrome rather than just DOM content.
Conditions required:
  • Victim follows social engineering prompts
  • The required Passwords UI is reachable in the current browsing context
  • The exact gesture sequence succeeds on the target build
Where this breaks in practice:
  • Email gateway, browser isolation, and user training all cut conversion on multi-step lures
  • Many users never visit password-management UI during routine browsing
  • Any change in browser state, language pack, or UX timing can break exploit reliability
Detection/coverage: Traditional vuln scanners cannot see this step; telemetry is mostly limited to browser history, URL clicks, help-desk reports, or advanced browser instrumentation.
STEP 03

Trigger the use-after-free in Passwords

If the gesture path is reached, the exploit attempts to hit a freed object in the Passwords component and turn that memory-safety flaw into controlled corruption. In theory this can end in code execution or comparable impact, which is why the vendor CVSS still looks scary on paper.
Conditions required:
  • The vulnerable Passwords code path is exercised
  • Heap state lines up well enough for a reliable trigger
  • The exploit survives Chrome hardening on the target build
Where this breaks in practice:
  • Use-after-free exploitation is reliability-sensitive, especially when the vector already requires high complexity
  • Modern Chromium mitigations and allocator behavior reduce one-shot exploit success
  • The bug is on macOS only, limiting scale compared with cross-platform Chrome bugs
Detection/coverage: Crash telemetry, EDR child-process anomalies, and browser crash dumps may catch failed attempts; signature-based IDS coverage is poor.
STEP 04

Convert corruption into meaningful impact

A successful exploit would need to turn the bug into code execution, data access, or other material security impact in the browser context. The title and supplied CVSS suggest high impact is possible, but there is no public sign that attackers have operationalized this path at scale.
Conditions required:
  • Exploit reliability is high enough on the victim's Mac
  • Post-corruption primitives are available
  • The attacker has a follow-on objective such as session theft or browser-context execution
Where this breaks in practice:
  • No KEV entry and no public campaign reporting reduce confidence in operational exploitation
  • No public PoC located as of 2026-06-05
  • A browser-context win is still less valuable than a clean sandbox escape unless chained
Detection/coverage: If exploitation succeeds, EDR and browser crash/telemetry pipelines are more likely to catch secondary behavior than the initial trigger.
03 · Intelligence Metadata

The supporting signals.

In-the-wild statusNo public evidence of active exploitation found as of 2026-06-05; not present in CISA KEV.
Proof-of-concept availabilityNo public PoC or weaponized GitHub repo located for this specific CVE as of 2026-06-05.
EPSS0.00068 from the supplied intel, which is very low and consistent with a niche, hard-to-reach exploit path.
KEV statusNot KEV-listed; current public KEV catalog review shows no entry for this CVE.
CVSS vector readoutCVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H means internet-reachable in theory, but high complexity plus required user interaction is the part that matters operationally.
Affected versionsSupplied intel says Google Chrome on Mac prior to 149.0.7827.53.
Fixed version149.0.7827.53 on Mac per the supplied intel; Chrome's early stable update for desktop also references 149.0.7827.53/.54 for Windows and Mac on 2026-05-29.
Exposure realityThis is a client-side browser flaw, so Shodan/Censys/FOFA exposure counts are not meaningful; the practical exposure metric is vulnerable managed Mac endpoints, not internet-facing hosts.
Disclosure timingSupplied intel says 2026-06-04. Public Chrome 149 early stable rollout for Mac/Windows began 2026-05-29.
Record confidenceThe affected version and release timing are consistent with Chrome's June 2026 release train, but I could not independently verify a public CVE record for CVE-2026-10900 by exact ID as of 2026-06-05; assessment is therefore based on the supplied intel plus official Chrome release context.
04 · The Call

noisgate verdict.

Final Verdict
DOWNGRADED to MEDIUM (5.6/10)

The single biggest downgrade factor is the required specific UI gesture sequence. That turns this from a broad drive-by browser bug into a narrow, user-steered exploit path on macOS only, which materially reduces both reachable population and attacker reliability.

MEDIUM Severity reassessment against the supplied title/vector
LOW Exact public CVE record verification for `CVE-2026-10900` by ID
HIGH Operational conclusion that UI-driven Mac-only browser bugs are less urgent than broad drive-by Chrome RCEs

Why this verdict

  • Start at the vendor's 7.5, then subtract for reachability: browser memory corruption is serious, but this one is not a plain one-click web RCE.
  • Specific UI gestures are a heavy friction point: this implies a multi-step lure and a narrow code path, not passive browsing.
  • Mac-only meaningfully trims fleet exposure: in most enterprises, Chrome-on-Mac is a subset of endpoints, not the full browser estate.
  • High complexity compounds the downgrade: reliable UAF exploitation already takes work; tying it to a finicky UI path makes it worse for the attacker.
  • No exploitation signal: no KEV, no public campaign reporting, and the supplied EPSS is extremely low.

Why not higher?

Because this is not broad population, low-friction exploitation. The combination of UI:R + AC:H + macOS-only + Passwords-specific path means several prerequisites must line up before the attacker even reaches the bug, and there is no public evidence that anyone has done that at scale.

Why not lower?

Because it is still browser memory corruption in a widely deployed product exposed to internet content. If the victim path is reached, the theoretical impact is still severe, and managed Macs used by admins, developers, or executives carry enough value that this should not be dismissed as harmless.

05 · Compensating Control

What to do — in priority order.

  1. Enforce Chrome auto-update on managed Macs — Use MDM or your endpoint management stack to force Chrome version compliance and relaunch behavior on macOS. For a MEDIUM noisgate verdict there is no mitigation SLA — go straight to the 365-day remediation window, but in practice browsers should be rolled much faster because update automation is low-cost.
  2. Reduce browser-UI lure success — Harden phishing controls, safe-browsing enforcement, and browser isolation for high-risk users so attackers struggle to walk users through a fragile click-sequence exploit. There is no mitigation SLA at this severity, so treat this as a risk-reduction measure while patching inside the 365-day remediation window.
  3. Limit privileged browsing from Macs — Keep admin sessions, production control-plane access, and sensitive SaaS administration out of the same browser profile used for general web activity. This does not fix the bug, but it cuts the value of a successful browser-context compromise while you remediate within the 365-day remediation window.
  4. Watch for stale Chrome on executive and developer Macs — Prioritize the subset of Mac endpoints that routinely hold high-value sessions, secrets, and password-manager data. Even with no mitigation SLA for MEDIUM, these users are where a niche browser exploit pays off the most.
What doesn't work
  • A network IPS signature will not save you here; the malicious traffic can look like ordinary browsing and the exploit depends on client-side state and UI interaction.
  • Perimeter scanning is mostly irrelevant because Chrome is a client application, not an exposed server service.
  • MFA alone does not materially stop exploitation of a local browser memory bug; it may limit downstream account abuse, but it does not prevent the trigger.
06 · Verification

Crowdsourced verification payload.

Run this on the target Mac endpoint locally, via SSH, or through your MDM script runner. Invoke it as bash check_cve_2026_10900_chrome_mac.sh "/Applications/Google Chrome.app"; no root is required, but the script must be able to read the app bundle's Info.plist.

noisgate-verify.sh
BASHREAD-ONLYSAFE
#!/bin/bash
# check_cve_2026_10900_chrome_mac.sh
# Purpose: Check whether Google Chrome on macOS is older than 149.0.7827.53
# Output: VULNERABLE / PATCHED / UNKNOWN
# Exit codes: 0=PATCHED, 1=VULNERABLE, 2=UNKNOWN

set -u

TARGET_VERSION="149.0.7827.53"
APP_PATH="${1:-/Applications/Google Chrome.app}"
PLIST_PATH="$APP_PATH/Contents/Info.plist"

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

if [ "$(uname -s)" != "Darwin" ]; then
  echo "UNKNOWN: This script is intended for macOS targets only."
  exit 2
fi

if [ ! -d "$APP_PATH" ]; then
  echo "UNKNOWN: Chrome app bundle not found at '$APP_PATH'."
  exit 2
fi

if [ ! -f "$PLIST_PATH" ]; then
  echo "UNKNOWN: Info.plist not found at '$PLIST_PATH'."
  exit 2
fi

if ! command -v /usr/bin/defaults >/dev/null 2>&1; then
  echo "UNKNOWN: macOS 'defaults' utility not available."
  exit 2
fi

INSTALLED_VERSION=$(/usr/bin/defaults read "$APP_PATH/Contents/Info" CFBundleShortVersionString 2>/dev/null)

if [ -z "${INSTALLED_VERSION:-}" ]; then
  echo "UNKNOWN: Could not read installed Chrome version."
  exit 2
fi

if version_lt "$INSTALLED_VERSION" "$TARGET_VERSION"; then
  echo "VULNERABLE: Google Chrome version $INSTALLED_VERSION is older than $TARGET_VERSION"
  exit 1
else
  echo "PATCHED: Google Chrome version $INSTALLED_VERSION is >= $TARGET_VERSION"
  exit 0
fi
07 · Bottom Line

If you remember one thing.

TL;DR
Monday morning: pull a Mac-only Chrome version report, identify endpoints still below 149.0.7827.53, and roll them through normal browser maintenance rather than declaring an incident. For a MEDIUM verdict there is noisgate mitigation SLA: no mitigation SLA — go straight to the 365-day remediation window; the noisgate remediation SLA is ≤ 365 days, but because this is Chrome and updates are operationally cheap, most enterprises should close it in the next routine browser wave rather than letting it age.

Sources

  1. Chrome Releases - Early Stable Update for Desktop (149.0.7827.53/.54)
  2. Chrome Releases - 2026 archive / release context
  3. CISA Known Exploited Vulnerabilities Catalog
  4. NVD Search
  5. NVD - CVE-2026-10000 (analogous recent Chrome Passwords flaw)
  6. Chrome CNA information at CVE.org
  7. Canadian Centre for Cyber Security advisory for Chrome 149.0.7827.53/54
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.