The critical feature that made Mosh for Chrome possible was the ability to send and receive UDP packets from within the browser. That is only available to Chrome Packaged Apps, which are not supported on mobile Chrome. You cannot implement a stand-alone mosh within a browser without this feature. At best you could implement a web frontend to some kind of mosh proxy that runs on a server. Such things already exist for ssh, but it would be an engineering challenge (to say the least) to make one for mosh that actually realized the specific benefits of mosh within the constraints of TCP connections. Plus you'd have to implement a lot of mosh in JavaScript (e.g, state sync protocol, predictive local echo, etc.); I was able to avoid reimplementing mosh by using Native Client, which is also not supported in mobile Chrome.
So in short, no I have no plans to make mosh work in the mobile browser. ;) I'm pretty happy using JuiceSSH on Android. If mobile Chrome begins to support Packaged Apps and Native Client, I would then consider seeing about making Mosh for Chrome work in it, but I don't know what plans there may be for that.
And to your point about quick deployment of patches, it really isn't any different between Chrome Apps and Android/iOS. They all enjoy an online "store" and automatic deployment of updates to the client.
And to your point about the complexity of testing on various mobile devices: The primary difference among devices is screen size, which would have to be tested even for a web app, so the developer would not be insulated from that responsibility. There are other complexities among devices, but an app this simple may not bump up against too many of them.
If you want to have an impact, I'd say your time would be best spent making a native iOS app. I don't know if it fits within their ecosystem, though; they're pretty strict about what apps are allowed to do.