What problem does it solve? Players of the Tywish on-chain lottery propose alternative winning rules in casual natural language, which a deterministic game engine cannot evaluate directly. This Skill interprets each proposed rule into one objective, measurable winning condition without changing the player's intent, and flags genuinely ambiguous rules for clarification. ## Core Features & Use Cases - Intent-Preserving Normalization: Rewrites a player's rule as a single measurable sentence referencing only the six-number guess and the six-number official draw, preserving quantity, property, position, and direction invariants. - Structured Clarification Flow: Detects missing dimensions such as endpoint inclusion, positional relation, or incomplete ticket values, and returns one to three targeted clarification questions instead of guessing. - Deterministic JSON Output Contract: Produces a fixed-schema JSON object with title, normalizedText, officialInterpretation, clarification fields, and a player-facing dealerMessage written in the language of the original rule. - Use Case: A player writes "vence quem acertar o primeiro numero do sorteio"; the Skill returns an English compiler-facing interpretation comparing position 1 of guess and draw, plus a Portuguese chat message confirming the consequence. ## Quick Start Interpret the player's proposed winning rule "the guess with the most drawn numbers wins" into the standard JSON interpretation contract.
