It is in the official extension repository [1]. I had to put it there because of how difficult/impossible it is to run a local extensions. But it's definitely not designed for general use. I'll be glad to guide you if the tiny amount of help I've included isn't enough.
The source [2] is available too.
The actual idea of the extension is "Style stuff using regular expressions". So you specify a regex; what to search, attributes and/or text (text can cause a lot of undesired results so I don't use it much); and css styles to add to the elements containing matches. I only really use it to apply visibility: hidden to any element with an attribute matching one big regex that's basically just a bunch of domains regex |'d together. The example in the options UI is what I'm still using years later, I've just added more domains.
EDIT: I just realize the example has "searchText" set to true, you'll probably want to set that to false.
To give you some idea of how user unfriendly it is, configuration is JSON in a textarea and the regexs are defined in a string in that JSON, necessitating the use of many backslashes for escaping from both regex and strings. It's basically brainfuck :)
I originally started with just taking out tags but that idea was quickly shot down. Lots of links aren't <a> tags. Sometimes the domain being linked to is buried in some JS in a onX event attribute, or some data-* attr that the UI framework uses and a bunch of other oddities. So the code searches all attrs. It's not super efficient but I've never noticed it causing a slow down.
Another problem is, often, the original request doesn't even contain the links, they're loaded after. So it uses a MutationObserver and looks at the new elements for matches.
Does it just strip out the <a> tag?