The affected boundary was between page and extension
According to the maintainer’s description, a web page could embed and influence the element-picker interface so that a user click created arbitrary trusted filters. The researcher explains that trusted filters can run JavaScript in other sites’ contexts, making this more consequential than merely hiding a page element. The official Firefox MV2 scenario depends on specific scriptlet-filter behavior involving web-accessible extension resources. It is therefore not an unconditional claim that opening any page automatically executes arbitrary code without interaction.
The version history is more precise than the headline
The advisory lists versions 1.75.2 and earlier as affected and 1.75.4 as patched. It dates the emailed report to November 1. An initial defensive change shipped in 1.75.2 on November 2, while verification of a secret token before launching the picker shipped in 1.75.4 on November 3. Additional filter validation and restrictions involving extension resources followed in 1.76.0 on November 8. Those separate dates matter: renewed attention on November 15 should not be presented as the first day users could obtain a fix.
Privileged interfaces need their own origin checks
The issue illustrates a trust-boundary failure when extension-privileged tooling accepts messages influenced by the surrounding page without adequately verifying their origin. The researcher describes a required user click and the possibility of concealing the relevant interface. The defensive changes strengthen launch verification and validation of the proposed filter. We did not execute the published demonstration or test exploitation against third-party websites. The relevant engineering lesson is the boundary between untrusted page content and privileged extension actions, supported by the maintainer’s documented remediation sequence.
Do not expand the advisory beyond its evidence
The researcher also mentions a related, less problematic issue involving uBO Lite; that discussion should remain distinct from the scope and version fields of the reviewed advisory. The official record labels severity High but does not provide a numeric CVSS score or known CVE. It also does not establish widespread exploitation or a specific theft of user data. For readers, the actionable distinction is their installed version and receipt of official updates, without assuming every uBlock variant is affected identically or treating a disclosed technique as evidence of a confirmed real-world breach.

Discussion.
Your perspective matters. Sign in, pick a nickname and choose your cyber identity.
Loading discussion…
Your account name, email and photo stay off your comments.