# uBlock Origin element picker one-click universal XSS (GHSA-cmqw-9r2w-6wv2) affecting Tor Browser on Tails — fixed in Tails 7.17.1

> A one-click universal XSS in uBlock Origin's website-driven element picker let a malicious page drive the picker in a hidden iframe and make the extension save a trusted scriptlet filter, running attacker JavaScript on any other site the user visits. Tails 7.17.1 (2026-11-17) temporarily removes uBlock Origin; upstream fixed it in uBlock Origin 1.75.4 and 1.76.0. No in-the-wild exploitation is known.

- **Published:** 2026-11-18T10:23:18Z
- **Last reviewed:** 2026-11-18T10:23:18Z
- **Canonical:** https://intel.threadlinqs.com/threat/TL-2026-6325
- **Card image:** https://intel.threadlinqs.com/og/threat/TL-2026-6325.png
- **ID:** TL-2026-6325
- **Severity:** HIGH
- **Category:** VULNERABILITY
- **Status:** PATCHED
- **Detections:** 9 (1 shown free below) · **IOCs:** 7

## Description

The uBlock Origin element picker UI (web_accessible_resources/epicker-ui.html) is loaded from a moz-extension:// origin and accepted postMessage/MessageChannel commands from any embedding page without verifying that the sender was the legitimate content script. A malicious website could (1) learn the per-install extension UUID by hooking XHR on a request made by a default uAssets filter (lootdest.*##+js(no-fetch-if, adsbygoogle, war:googlesyndication_adsbygoogle.js)) that fetches an extension web-accessible resource; (2) embed the picker in an iframe; (3) send messages that start the picker and set a filter, where the 'opt' field is appended to the filter so a newline allows injection of an arbitrary filter, including trusted scriptlet filters such as trusted-rpnt, which the picker normally refuses; and (4) obtain one user click on the picker's 'Create' button, which can be achieved by making the iframe invisible (opacity 0) and overlaying it under the cursor (the researcher's demo did not implement this step). Filters added through the picker are treated as trusted until a recompile, so the attacker can execute JavaScript in the context of any domain, e.g. webmail, reading or altering visible content.

The advisory (GHSA-cmqw-9r2w-6wv2, High, no CVE, no CWE listed) affects uBlock Origin 1.75.2 and lower on the Firefox MV2 build; the MV3 uBO Lite had a similar but less severe issue. The reporter (syvb / Smitty van Bodegom) emailed the maintainer on 2026-11-01. Fixes: 1.75.2 (2026-11-02) avoids tainted XHR methods in scriptlets; 1.75.4 (2026-11-03) launches the picker only with a verified secret token; 1.76.0 (2026-11-08) validates the filter created in the picker and blocks war:-related filters. The reporter noted the 1.75.4 fix could be bypassed by redefining the XHR responseText getter to leak the UUID, addressed by the 1.76.0 filter validation. Coverage recommends 1.76.4 (2026-11-15) as the current stable.

In Tor Browser on Tails, the Standard and Safer security levels are affected; Safest is not. Tails 7.17.1 temporarily removes uBlock Origin to remove the risk and plans to restore it later. Restarting Tails clears any malicious uBlock Origin configuration injected, but does not undo data theft that already occurred. Tails is not aware of attacks using the flaw. Severity HIGH is an analyst assessment; no CVSS or CVE is stated by sources.

## MITRE ATT&CK

- T1659 Content Injection
- T1204.001 User Execution: Malicious Link
- T1059.007 Command and Scripting Interpreter: JavaScript
- T1176.001 Software Extensions: Browser Extensions
- T1185 Browser Session Hijacking
- T1565.001 Data Manipulation: Stored Data Manipulation

## Free detection (Splunk SPL, Microsoft KQL, Sigma)

**uBlock Origin element-picker uXSS (GHSA-cmqw-9r2w-6wv2) — Navigation to lootdest.* Exploit-Host Pattern** (`TL-2026-6325-SPL-001`)

The public uBlock Origin uXSS PoC (GHSA-cmqw-9r2w-6wv2, no CVE; Tails 7.17.1 fix) leaks the per-install extension UUID by hooking the XHR made when the default uAssets filter lootdest.*##+js(no-fetch-if, adsbygoogle, war:googlesyndication_adsbygoogle.js) fires, so the attacker page must be served from a hostname whose label is lootdest under any TLD. This search flags web-proxy requests to lootdest.* hosts, the network-visible precondition of the one-click chain (drive-by delivery, user-opened link, content injection into the extension). It is an exposure/precondition signal, not proof of compromise: pivot to SPL-002 for the post-exploitation storage write.

- **Languages:** Splunk SPL, Microsoft KQL, Sigma
- **ATT&CK:** T1189, T1204.001, T1659
- **Data sources:** web_proxy_squid_access, proxy
- **Severity:** medium
- **Confidence:** medium

### Splunk SPL

```spl
index=proxy sourcetype="squid:access" url=*lootdest*
| rex field=url "^(?:[a-z]+://)?(?<req_host>[^/:?]+)"
| eval req_host=lower(req_host)
| where match(req_host, "(^|\\.)lootdest\\.[a-z0-9.-]+$")
| stats count AS requests dc(url) AS distinct_urls min(_time) AS first_seen max(_time) AS last_seen values(url) AS sample_urls BY src, user, req_host
| convert ctime(first_seen) ctime(last_seen)
| sort - requests
```

### Microsoft KQL

```kql
// TL-2026-6325 uBlock Origin element-picker uXSS: lootdest.* exploit-host precondition
DeviceNetworkEvents
| where Timestamp > ago(24h)
| where ActionType in ("ConnectionSuccess", "ConnectionRequest", "HttpConnectionInspected")
| where RemoteUrl matches regex @"(?i)(^|\.)lootdest\.[a-z0-9.-]+$" or RemoteUrl matches regex @"(?i)^https?://([a-z0-9-]+\.)*lootdest\."
| where InitiatingProcessFileName in~ ("firefox.exe", "firefox", "firefox-esr", "tor-browser", "tor.exe")
| summarize Requests=count(), FirstSeen=min(Timestamp), LastSeen=max(Timestamp) by DeviceName, InitiatingProcessAccountName, InitiatingProcessFileName, RemoteUrl
| order by Requests desc
```

### Sigma

```yaml
title: uBlock Origin uXSS Exploit Host Pattern - Proxy Request to lootdest.* Host
id: 6325a001-0001-4a6b-8c01-000000006325
status: experimental
description: Detects web proxy requests to hosts with a lootdest label, the hostname pattern needed by the public GHSA-cmqw-9r2w-6wv2 uBlock Origin element-picker uXSS proof of concept to leak the extension UUID via the default uAssets lootdest.* scriptlet filter.
references:
    - https://github.com/gorhill/uBlock/security/advisories/GHSA-cmqw-9r2w-6wv2
    - https://iter.ca/post/ubo-uxss/
    - https://blog.torproject.org/new-release-tails-7_17_1/
author: Threadlinqs Team
date: 2026-11-18
tags:
    - attack.initial-access
    - attack.t1189
    - attack.execution
    - attack.t1204.001
    - attack.t1659
logsource:
    category: proxy
detection:
    selection:
        cs-host|re: '(?i)(^|\.)lootdest\.[a-z0-9.-]+$'
    condition: selection
falsepositives:
    - Legitimate visits to real lootdest link-shortener sites
    - Security research and lab testing of the public PoC
level: medium

```

**Known false positives:**

- Users legitimately browsing real lootdest.com / lootdest.org link-shortener sites
- Authorised red-team or lab reproduction of the public PoC

**Tuning notes:** Allow-list the genuine lootdest registrable domains seen in your estate and alert only on new lootdest TLD variants or on hosts first seen in the last 7 days. If your proxy logs the User-Agent, restrict to Firefox/Tor Browser. Pair with SPL-002 and the vulnerable-version inventory (KQL-001) to prioritise endpoints. Squid logs CONNECT host:port for HTTPS, so the host regex is applied to the extracted hostname.

### 8 more detection rules (Splunk SPL, Microsoft KQL, Sigma), unlock with TL Creds

- Vulnerable uBlock Origin (Firefox MV2) Installed — Exposed to GHSA-cmqw-9r2w-6wv2 Element-Picker uXSS (locked: unlock with TL Creds)
- uBlock Origin Storage Written Inside Tor Browser Profile (Tails/Linux and Windows) — Element-Picker uXSS Persistence (locked: unlock with TL Creds)
- Bulk Mailbox Reads via OWA Browser Session — Possible Injected-Script Webmail Collection (T1185) (locked: unlock with TL Creds)
- uBlock Origin uXSS Precursor — Browser DNS Resolution of lootdest.* Hosts (locked: unlock with TL Creds)
- Webmail Inbox Forwarding Rule Created Following Possible uXSS Session Abuse (T1185 / T1114.003) (locked: unlock with TL Creds)
- Firefox/Tor Browser Profile Extension (.xpi) Dropped by Non-Browser Process — Browser Extension Persistence (locked: unlock with TL Creds)
- uBlock Origin Extension Storage Written by Firefox/Tor Browser — Trusted Scriptlet Filter Persistence (trusted-rpnt) (locked: unlock with TL Creds)
- Webmail-Origin POST to Unfamiliar Destination — Injected JavaScript Exfiltration After uBlock Origin uXSS (locked: unlock with TL Creds)

Sign up free to unlock the detection rules: https://threadlinqs.com/?modal=signup

## Sources

- [Tails 7.17.1 release notes (Tor Project blog)](https://blog.torproject.org/new-release-tails-7_17_1/)
- [Universal XSS in uBlock Origin from website-driven element picker (researcher write-up)](https://iter.ca/post/ubo-uxss/)
- [GitHub Security Advisory GHSA-cmqw-9r2w-6wv2](https://github.com/gorhill/uBlock/security/advisories/GHSA-cmqw-9r2w-6wv2)
- [Cyber Security News: Tails and Tor Browser update to fix uBlock Origin flaw](https://cybersecuritynews.com/tails-tor-browser-update-to-fix-ublock-origin-flaw/)
- [AllAboutCookies: One-click uBlock Origin bug could let malicious sites run code on any page](https://allaboutcookies.org/ublock-origin-security-flaw-update)
- [Privacy Guides discussion: Universal XSS in uBlock Origin from website-driven element picker](https://discuss.privacyguides.net/t/universal-xss-in-ublock-origin-from-website-driven-element-picker/42249)
- [uBlock Origin element picker documentation](https://github.com/uBlockOrigin/uBlock-issues/wiki/Element-picker)

## Full data

The other 8 detection rules (Splunk SPL / Microsoft KQL / Sigma) unlock with TL Creds; a free account includes daily creds: https://threadlinqs.com/?modal=signup
Indicator values unlock with TL Creds, at a lower price on Red and Purple; AI agents can query rules and IOCs through the Threadlinqs MCP server (Purple): https://intel.threadlinqs.com/mcp

Canonical: https://intel.threadlinqs.com/threat/TL-2026-6325
