Critical gaps across all 8 tools. No input schemas visible in provided source code, cannot verify parameter types, constraints, or structure. Tool descriptions are present but generic (10-50 chars), lacking WHAT/WHEN/WHY context required for LLM selection. No output schemas documented. No error handling guidance. Tools appear to be registered via mount_tools() in server.py, but actual schema definitions are not shown in the provided code sample, making it impossible to verify compliance with JSON Schema or parameter documentation standards. This is a HARD CAP scenario: without visible schemas, schema scores must be 0.

• Add explicit JSON Schema input definitions to each tool with type, description, and constraints for every parameter. Example: todos tool should define parameters like 'action' (enum: create|update|delete|list), 'task_id' (string, UUID format), 'title' (string, 1-256 chars), 'status' (enum: pending|in_progress|completed).
• Expand tool descriptions to 50-200 characters following pattern: 'WHAT: [action]. WHEN: [use case]. RETURNS: [key fields].' Example: 'Create, update, or list todo items. Use when tracking tasks or project work. Returns task_id, title, status, created_at.'
• Document output schemas for each tool showing field names, types, and pagination. Example: todos list returns {items: [{id, title, status, created_at}], total_count, next_cursor}.
• Add parameter descriptions explaining format, constraints, and dependencies. Example: 'task_id (string, UUID format): Unique identifier of the todo item. Required for update/delete operations.'
• Implement error handling that returns actionable guidance: 'Task not found (id=xyz). Try list_todos() to find available tasks.' Include error classification (retryable|user_fixable|fatal).