What problem does it solve? Quantitative model research involves literature review, model specification, training component integration, and risk handoff, all of which must follow organizational authorization and traceability rules. This Skill defines the model group's task workflow so that every step uses verified capabilities, shared memory, and authoritative training components instead of unverified assumptions. ## Core Features & Use Cases - Capability-First Execution: Queries list_capabilities and search_memory before acting, verifying the real integration status of models, features, labels, and training components. - ModelSpec Contract Output: Produces schemas.model.ModelSpec with fields like model_name, model_type, training data ranges, hyperparameters, feature dependencies, and risk metadata, explicitly flagging missing values instead of using mock data. - Governed Human Gates: Uses organization_reuse and organization_solution for coverage decisions and plan approval, with exact merge/permission approval for shared writes. - Use Case: A model group member starts a task to define a new factor model; the Skill guides them to check component availability, load the model-lit-review sub-skill for literature work, call authoritative training and OOS components, and deliver a traceable ModelSpec for risk team review. ## Quick Start Start a new model group task and ask the agent to verify available training components, review relevant literature, and produce a ModelSpec for the proposed model.
