Sorry for the trouble! Should be fixed in v1.3.1 now!

No true! It is actively *maintained*, I'm just unlikely to add new features without sufficient motivation at this point (maintenance ≠ feature development). ;-)
(Obviously I will accept patches for any of the requested features, so feel free donate your time if money isn't available to you!)

Thank you for your review! The “huge cache of hundreds (!!!) of files […], which may be completely removed” are in fact used to generate sensible values for the `navigator.*` browser identification values. Parsing information like operating system, CPU type, device type, … is not a simple affair as there are no standards at all – one just has to guess based on the COMMENT parts of the User-Agent string. Fortunately the BrowsCap project has maintained a list of matching patterns for figuring out this information since 1997 that is extremely reliable. As such, I will continue using it even if it means bloating the extension size by about 3MiB. The fact that it allows everything to just work out-of-the-box with just a single input value (the User-Agent string itself) definitely outweighs that, by comparison small, price.
I would like to think that, if you don't need this complete level of emulation, then you may use just about any of the other User-Agent switching extensions because – to my knowledge this capability is unique to this extension.
- Alexander

Thanks for the review!

Could you please elaborte what you mean by “The downside is adding or changing them does not work”? I do not unfortunately understand what you mean by this.

Having the CLI tools options available is good for testing for the exact behaviour you describe as “bad”. Some sites also hide/modify content from web browsers, but only deliver them to bots and specialized client software for automated processing – feeds/podcasts are a prominent example of this.