Case Retrieval is a skill that provides a structured interface to the Deli Legal (得理) Open Platform API, giving AI agents access to Chinese judicial documents including judgments, rulings, and mediation documents. It is built for legal practitioners and developers constructing legal AI agents that need accurate, live case data rather than relying on a language model’s training knowledge.
The skill supports two search modes: precise keyword filtering by case number, court name, or cause of action, and long-text semantic matching that maps natural language descriptions of case facts to professional legal vectors. This makes it suited for similar-case matching (类案检索), a task that requires domain-specific vectorization of Chinese legal concepts.
Once cases are retrieved, the skill can automatically extract core facts, identify disputed issues, and surface key judicial reasoning from lengthy documents. For multi-case research, it compares several cases to synthesize consistent judicial rules and decision-making patterns — a process referred to as rule induction.
Output is delivered as structured Markdown: case list tables, detailed case profiles that include party information and trial procedure details, and comparative analysis grids for multi-case work. The skill enforces strict data-fetching constraints to keep legal citations accurate and complete.
This skill is not appropriate for legal systems outside China, for non-judicial document types, or for workflows that do not require access to the Deli Legal API.