Six skills for working with SpiceGrinder, Obsvra's local, deterministic synthetic-data engine: draft a model from a description, fit one to a real data sample, explain an existing model, write custom generators and filters, audit a model for PII risk, and profile it for performance.
Grounded in the real component catalog.
Skills 6 in the Claude listing
Build spicegrinder model Given a natural language description, create a valid, runnable SpiceGrinder model. Explain model A user provides a SpiceGrinder model and asks for an explanation of what the model is meant to simulate, or appears unsure of what a model should represent. Extend spicegrinder Helps write, edit, or debug custom SpiceGrinder generators, filters, and business objects -- the IGenerator/IFilter interfaces, @ComponentInfo/@Parameter annotations, and the Create() factory pattern Fit model to sample Analyzes a real data sample (a CSV export, database dump, or other tabular data) and produces a matching SpiceGrinder model definition by statistically fitting each column to SpiceGrinder's available Pii audit A user has a SpiceGrinder model file and wants to check for any potential PII issues, or indicates via conversation that the data generated by a model might be PII-adjacent. Tune model performance A user provides a SpiceGrinder model and wants a performance review -- identifying inefficient structure, likely bottleneck nodes, and suggestions for improvement based on the Model Analyzer's documen