I'm impressed with the one year time frame you give yourself. I wish I had that kind of patience for a project.
Choosing to do it in Construct is probably not the fastest method, but it makes for a challenge.
I know you already plan on supporting a subset of the plugins but I'd recommend only implementing only a subset of the engine's features, at least initially. The idea is to get something working early on so you can get motivation and/or a better picture of the work involved early on. This will lead to needing to refactor your code later as you add stuff, but you'll be doing that anyway if you tried to implement everything at once.
You may find it interesting that events aren't converted to javascript. They are run by the event engine which in turn runs the javascript in plugins and the engine, while it also updates the picked object lists.
Anyways I'm sure it will be very good coding practice no matter the result.