Threat reportVulnerabilityTL-2026-3787

CrossBarking: Opera Browser 0-Day Private API Abuse via Cross-Browser Extension Store Attack

highPATCHED

CrossBarking (TL-2026-3787), also tracked as CrossBarking, is a high-severity software vulnerability, first published 2026-10-16. It has no confirmed attribution, affects Opera Software Opera browser, maps to 8 MITRE ATT&CK techniques (T1036, T1059.007, T1113), and is covered by 9 detection rules and 13 indicators of compromise.

Severity
HIGHAssessed severity
CVEs
0None referenced
Techniques
8MITRE ATT&CK
Actors
0Not attributed
Detection rules
9SPL · KQL · Sigma
IOCs
13Indicators of compromise

Key facts for TL-2026-3787

Threat ID
TL-2026-3787
Also known as
CrossBarking
Severity
HIGH
Status
PATCHED
Category
VULNERABILITY
First published
Last reviewed
Attribution confidence
LOW
Motivation
UNKNOWN
Target sectors
technology, consumer, enterprise
Target regions
Global
Detection rules
9
Indicators of compromise
13

Malware and tooling in CrossBarking

Malware and tooling: Guardio 'puppy' PoC extension

How CrossBarking works

Guardio Labs disclosed 'CrossBarking', a flaw in Opera that let a malicious extension (e.g. one distributed via the Chrome Web Store) inject content scripts into hardcoded privileged Opera domains and reach permissive Private APIs. This enabled screen capture, browser setting modification (PoC: DNS-over-HTTPS hijack) and account hijacking. Opera patched it on 2024-09-24; neither Opera nor Guardio reports in-the-wild exploitation.

CrossBarking abuses the way Opera exposes browser 'Private APIs' to a set of hardcoded, privileged web origins that back first-party features such as Pinboard and Opera Wallet. Guardio Labs found that these origins included Opera-owned subdomains and test hosts (crypto-corner.op-test.net, op-test.net, pinboard.opera.com, gxc.gg, opera.atlassian.net) and third-party integration domains (vk, Instagram, Yandex). Opera did not block extension content scripts on them. An extension that declares a content script matching those domains (the PoC used <all_urls>) can inject script into the page's execution context and call Private APIs that ordinary web content and ordinary extensions cannot reach.

The delivery path is a cross-store attack. Opera's own add-ons store is manually reviewed, but Opera redirects users who look for some extensions to the Chrome Web Store. Guardio published a benign-looking 'puppy' extension there (justification: 'Add cute puppies to every page') as an unlisted item. It was approved within about 24 hours, because Chrome's policy checks look for Chrome-specific capabilities and an Opera-specific exploit triggers no alarms. The victim must install it in Opera and dismiss Opera's warning about third-party-store extensions.

Guardio's PoC built the payload as a base64-encoded JSON blob in a URL fragment on the privileged host (https://crypto-corner.op-test.net/#puppiesOn-<base64>). The script was injected dynamically through DOM manipulation, which sidesteps page CSP because extension-injected scripts are treated differently from inline scripts. The blob called chrome.settingsPrivate.setPref() to set dns_over_https.mode=secure and dns_over_https.templates to the PoC resolver https://bad.dnsserver.xyz, redirecting all browser DNS queries to an attacker-controlled DoH server for adversary-in-the-middle, phishing and surveillance. Guardio notes the exploit could run from the extension's onInstall event with no further user interaction.

Other exposed Private APIs per Guardio: pinboardPrivate (screen capture of tabs), chrome.cookies (session cookie theft and account hijacking) and management/addonsPrivate (disable or install extensions, including security tools). Fix, deployed 2024-09-24: content scripting was disabled on high-permission domains (similar to the Chrome Web Store domain protection) and third-party domain privileges (VK, Instagram, Yandex) were removed. Private APIs remain reachable in those contexts, so XSS on a privileged origin stays a valid vector, and Opera has started refactoring to remove the Private API architecture. No CVE or CVSS was assigned, affected Opera versions were not specified, and Opera and Guardio state there is no evidence of in-the-wild exploitation. The hostnames listed here are legitimate Opera infrastructure abused by design, so they are context for hunting extension behavior, not block-list candidates.

MITRE ATT&CK techniques used in TL-2026-3787

Defense Evasion

T1036 Masquerading

Execution

T1059.007 JavaScript

Collection

T1113 Screen Capture; T1185 Browser Session Hijacking

Persistence

T1176.001 Browser Extensions

Credential Access

T1539 Steal Web Session Cookie; T1557 Adversary-in-the-Middle

defense-impairment

T1685 Disable or Modify Tools

Affected products and versions in CrossBarking

  • Opera Software — Opera browser
    Vulnerable versions: Versions not specified in sources (pre-2024-09-24 builds)
    Fixed in: Fix deployed 2024-09-24

Remediation for CrossBarking

Patches

  • Opera server/browser-side fix of 2024-09-24: content scripting disabled on high-permission domains; third-party domain privileges (VK, Instagram, Yandex) removed

Immediate actions

  • Update Opera to the current release (fix deployed 2024-09-24)
  • Audit extensions installed in Opera that came from the Chrome Web Store and remove any not required
  • Review Opera DNS-over-HTTPS settings (dns_over_https.mode / dns_over_https.templates) for unexpected resolvers

Workarounds

  • Do not dismiss Opera's third-party-store extension warning
  • Block or alert on extensions requesting <all_urls> content scripts from unknown publishers

Longer-term hardening

  • Restrict extension installation in managed browsers to an allow-list
  • Monitor endpoint and DNS telemetry for browsers using unapproved DoH resolvers
  • Prefer Opera's manually reviewed add-ons store over third-party stores

Weaknesses (CWE) in CrossBarking

CWE-269, CWE-284

Timeline of CrossBarking

  • Opera deploys the fix: content scripting disabled on high-permission domains and third-party domain privileges (VK, Instagram, Yandex) removed.
  • As of publication Opera has started refactoring to eliminate the vulnerable Private API architecture; XSS on privileged origins remains a theoretical vector.
  • The Hacker News reports the fix and lists privileged domains with Private API access (crypto-corner.op-test.net, op-test.net, gxc.gg, opera.atlassian.net, pinboard.opera.com, instagram.com, yandex.com).
  • Opera and Guardio state there is no evidence of in-the-wild exploitation of this scenario.
  • Opera publishes its security blog post crediting Guardio and stating no Opera user was affected before the patch.
  • Guardio Labs publishes 'CrossBarking' research describing the cross-browser extension store attack and PoC (DoH hijack via settingsPrivate.setPref).
  • CyberInsider publishes coverage of the cross-browser attack; press reports confirm the PoC extension was published on the Chrome Web Store and that no exploitation was known before the patch.

Sources cited for CrossBarking

Detection coverage for TL-2026-3787

As of 2026-10-16, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-3787 across Splunk SPL, Microsoft KQL and Sigma, covering 13 indicator(s) of compromise. The whole corpus is readable without an account; a free account unlocks full detection query text in Splunk SPL, Microsoft KQL and Sigma; paid tiers add raw indicator values, correlation and the MCP server. Threadlinqs MCP server · View plans.

9 detection rules (Splunk SPL, Microsoft KQL, Sigma) · Blue and above. Compare plans
13 indicators of compromise · Red and above. Compare plans

Threadlinqs Intelligence — Real-Time Threat Detection Platform

[ 0 threats ] [ 0 det ] [ CRIT: 0 ] [ HIGH: 0 ]
// threat_feed
$ sort --newest
Showing all threats

Live intelligence console

Threat level
Fig. 01 · Threat weatherIndexing the archive…
1 square = 1 threat · click to open

Latest Threats