Three tools with clear names and detailed descriptions, but significant schema and parameter documentation gaps. Tool names follow verb_noun pattern (list_, authorize, gate_history_get). Descriptions are substantive (200-400 chars) and explain purpose, prerequisites, and side effects. However, input schemas lack proper JSON Schema structure with explicit type declarations and parameter descriptions. The 'evidence' parameter in authorize is documented as 'object' with minimal guidance on expected fields. Output schemas are not formally documented. Error handling is present but lacks recovery guidance. Security considerations (audit trails, permission gates) are well-thought-out in implementation but not surfaced in tool definitions.

• Add formal JSON Schema definitions for all input parameters. For 'evidence' in authorize, document expected fields (e.g., 'evidence: {type: object, properties: {field1: {type: string, description: "..."}, ...}, required: [...]}'). Reference preconditions/<case>.py docstrings in the schema.
• Document output schemas explicitly. Create a schema definition for authorize's return object listing all fields (gate_state, cq_scores, reason_code, schema_version, etc.) with types and descriptions.
• Add parameter descriptions to all tools. For 'case', list valid values or explain how to discover them. For 'precondition_fn', state 'Must be a value returned by list_precondition_functions for the same case.'
• Enhance error messages with recovery guidance. When raising ValueError for unknown precondition_fn, include 'Call list_precondition_functions(case="...") first to see available options.' in the error message.
• Add pagination to gate_history_get. Introduce optional 'limit' (default 50, max 1000) and 'offset' parameters. Return total_count and next_offset in response to enable cursor-based pagination for large audit histories.