It sounds like you got it working. The SQLite implementation was my pet project about a year ago for some related work we are doing with my research. Our needs for the project were fairly simple, and in many ways the wrapper was a proof of concept to explore whether a working relational database could be done at all on a cRIO platform.
As you have probably learned, there is no "special source" used for compilation, though there are a couple of minor changes that need to be made in order for SQLite to compile correctly using the gcc toolchain. In principle, modifying new versions of SQLite to compile is a fairly straightforward process - I was able to get 3.6.18 to compile last night in a matter of minutes. I haven't, however, had the time to fully test the stability of such compiles in order to feel comfortable distributing them out to the community at large.
I am attaching to this post the source and makefiles I used to build the sqlite3.out files Brian posted for VxWorks. I have included two versions within this, both based on the 3.6.3 source. The first version, 3.6.3_base, has been extensively tested on a working prototype for several months and has (when used with the wrapper correctly) not resulted in a crash. I feel fairly confident in its stability, though I would recommend testing it quite a bit in your own application before putting it in a mission critical environment. The second version, 3.6.3_extended, adds some additional functionality (the vacuum function, for instance) which might be useful for various database applications, but has not been tested and verified.
With regard to the wrapper function, it was not intended to be the sum of all SQLite functions, but rather the core set of functionality needed for the proof of concept I was working on at the time (essentially select and non-select queries). I would certainly be more than happy to help if you have any questions regarding extending the wrapper or adding functionality to it.