A minimum spanning tree algorithm is implemented as part of the MST parser in OpenCog, here:
https://github.com/opencog/opencog/blob/master/opencog/nlp/learn/mst-parser.scm
The MST algo that's implemented is Boruvka's (not Prim or Kruskal) since it gives better results for natural-language parsing (speed and scalability are not important for this app.) The implementation is custom-tailored to text-parsing, and is not general purpose. It can't possibly be hard to generalize, right? :-)
All OpenCog code is licensed under the GNU AFFERO GENERAL PUBLIC LICENSE Version 3, 19 November 2007