nuke-dialog

Nuke cookie / GDPR dialogs on web pages

Concept

The concept here is that the web is full of un-desireable dialogs that one must - with fine motor control - click "I agree, consent, yes master", etc to dismiss.

Most of the logic employed in nuke-dialog is via a reverse css lookup where we find the selectors to target via the css rules themselves.

The secondary mechanism for discovery is via a MutationObserver that looks for nodes that change: class and attribute changes or new nodes that are created wit z-index > 0;

This is still naive as it depends on the textContent of cookie to exist in the nodes that we are targeting via the css or mutations.

TODOs

• Package index.js in to browser extensions
• Keep track of the modals dismissed locally in IndexDB, check the cache for what to dismiss before running the heavy-ish script that potentially does XHRs etc
• "Mark modal" functionality: allow users to right click to mark a modal DOM node manually.
• Discover modals that do not have z-index style rules. Ideas here will be great to hear about.
• I18N: The miniscule UI and string testing should be localized as needed.

MAYBEs

• Test for local IPFS gateway to post a newly-discovered modal CSS Selector and URL (or perhaps just the domain name) to a CRDT DB
• When the browser starts each URL visited is compared to the CRDT to be able to automatically discover the modals and dismiss them automatically.

ML model?

Is there a machine learning model that could be applied to the DOM to discover dialogs or other un-desired UX in web pages?