The KB is definitely json-ld (rdf graph representations) and PFP inputs and outputs are also meant to use much of the same terms. A consistent json-ld context helps with this.
Also, consistency of formats and styles (e.g. property names) across artifacts would help. This includes ontology term and label naming and consistent use of utility vocabularies (e.g. schema.org)
The KB is definitely json-ld (rdf graph representations) and PFP inputs and outputs are also meant to use much of the same terms. A consistent json-ld context helps with this.
Also, consistency of formats and styles (e.g. property names) across artifacts would help. This includes ontology term and label naming and consistent use of utility vocabularies (e.g. schema.org)