Hardens misleading-link detection for Amazon S3 and other shared-hosting services. The extension now compares tenant and bucket identities, recognizes additional destination cues such as secure sign-in and account-access text, detects brand impersonation on shared hosts, and limits visible-domain parsing work. Permissions, data practices, and local-only analysis are unchanged.

Reduces false warnings on legitimate Microsoft Teams government-cloud meeting and dial-in links, aka.ms help links, and Microsoft Safe Links-wrapped email URLs. Safe Links are unwrapped locally only on Microsoft's documented hostname, and their embedded destinations remain fully checked for suspicious patterns.

Reduces false warnings on legitimate Reddit links and rich link cards. Reddit's canonical, short, media, and deep-link hosts are recognized as related destinations, while unrelated app-link tenants and misleading visible-domain claims continue to be flagged. Adds broader automated and real-Firefox regression coverage.

Reduces false warnings on legitimate Google, cloud infrastructure, filenames, versions, dates, and other dotted text. Improves detection of misleading domains, shared-hosting boundaries, Unicode brand lookalikes, accessibility labels, dynamically added links, and subframe links. Warning labels remain isolated from conflicting webpage CSS.

Keeps warning labels upright and readable when webpages apply conflicting CSS. Adds Firefox ESR 140 desktop compatibility while retaining Firefox 142 as the Android minimum.