What problem does it solve? Contest problem statements for mathematical modeling competitions like CUMCM are dense, and missing a single condition, definition, or cross-question dependency can invalidate an entire solution. This Skill performs a sentence-by-sentence semantic translation of the full problem into precise modeling language, exposing hidden constraints and producing an auditable coverage ledger. ## Core Features & Use Cases - Sentence-by-sentence translation ledger: Preserves every exact source sentence with explicit facts, implied modeling signals, cross-sentence links, and omission-risk notes. - Mandatory Mermaid cross-question flowchart: Maps shared definitions and data through every numbered question, including parallel branches, convergence, and validation return edges. - Completeness audit: Verifies that every substantive source unit appears exactly once, every numbered question is covered, and pre-title boilerplate is excluded. - Use Case: A team receives a 2024 CUMCM Problem B PDF. They paste the full text and receive one UTF-8 Markdown report with an overview, terminology table, per-question input-task-output matrix, Mermaid dependency graph, and a list of the most easily misread sentences. ## Quick Start Use the bzd-problem-translator skill to translate this complete CUMCM problem statement sentence by sentence and generate the Markdown report with the Mermaid cross-question flowchart.