I looked into your code and here is what I found
Running HTML
This works fine if you make a couple small changes. 1. Your tests setting in the chutzpah.json file has a trailing comma. This exposes a bug where the deserializer returns this as a null entry and chutzpah isnt handling it. I will fix this issue but you can just delete the extra comma. 2. Once I deleted that command I saw that your chutzpah.json file has a tests setting that is listing just the .js file. But when running the html file then that is your test file. So your chutzpah.json is filter it out. If you update the tests setting to point to the html file this works.
Once I made those changes the test ran successfully.
Running Chutzpah.json
Looking into this the way you load the example.spec in SpecRunner you reference the .js extension in your require call. This tells require.js to ignore the baseurl and look at the file relative to current location. When you have chutzpah generate the test harness it generates a require call without extension. This means that file needs to be able to be loaded given the current baseurl setting.
Now there is a bug in chutzpah where it doesn't understand baseurls. See https://chutzpah.codeplex.com/workitem/214. The planned solution is to allow you to tell chutzpah in the chutzpah.json file what your baseurl is. Without that it is still possible to get this to work but you would need to tell chutzpah to place its generated .html file where your baseurl would point.
Let me know if that helps.