When using other analysis tools I often take advantage of assertions. For example using phantomas one can assert that the analysis "failed" when there are more than X number of HTTP requests total, or the CSS exceeds a certain number of bytes, etc.

Using assertions within Parker could be a nice way to control the growth of a CSS codebase. I'm looking at stats like "Total Important Keywords" or "Total ID selectors" — things like that.

Possible cons: it might get misused and people might be lured into a false sense of security (or even write code in weird ways to avoid making a certain stat rise).

Would this feature be worthwhile to anyone else? Are there other cons I'm overlooking?