What problem does it solve? Writing actuarial models in gaspatchio requires knowing the correct ActuarialFrame idioms, method signatures, and performance rules; guessing them leads to silent wrong results, broken lazy execution, and slow code. This Skill enforces a mandatory documentation lookup gate, model classification, and a three-phase build pattern so models are auditable, vectorized, and correct. ## Core Features & Use Cases - Mandatory doc lookup gate: Requires running uv run gspio docs "<method>" before using any unverified gaspatchio method, eliminating the ~70% error rate from guessed signatures. - Model classification: Routes models into standard projection, linear recurrence (accumulate), state-machine rollforward, fund aggregate, or cash flow waterfall classes before any code is written. - Three-phase build pattern with performance rules: Enforces setup/timeline/calculation phases, bans map_elements and Python loops, and provides a gotcha reference table of real production failures. - Use Case: When converting an Excel variable annuity model to gaspatchio, use this Skill to classify the model, look up Table.lookup and cumulative_survival signatures, build the projection timeline, and validate each section with run-single-policy. ## Quick Start Use the gaspatchio-model-building skill to write a monthly mortality projection model from my model_points.parquet file, looking up each method with gspio docs first.
