Four tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Tool names follow verb_noun convention (list_, generate, chat, embed). Descriptions are present but brief (10-50 chars), below the 194-char baseline for production tools. Input schemas are visible and typed, but parameter descriptions are minimal or missing context. No output schemas documented. Error handling is present but not recovery-focused. The 'messages' parameter in chat tool lacks detailed structure documentation (only states 'each with role and content' without specifying required fields or valid role values).

• Document output schemas for all four tools. Example: list_models should return {models: [{name: string, size: int64, modified_at: string}], total: int}. generate and chat should return {text: string, prompt_eval_count: int, eval_count: int}. embed should return {embeddings: number[][], dimensions: int}.
• Expand 'messages' parameter description in chat tool to: 'Array of message objects. Each must have: role (enum: system, user, assistant) and content (string). Optional: thinking (string, for models that emit reasoning). Example: [{role: "system", content: "You are helpful"}, {role: "user", content: "Hello"}]'.
• Add enum constraints for 'model' parameter. Instead of free-form string, call list_models internally or document: 'Model name must match one returned by list_models. Common examples: llama3:8b, qwen3:14b, nomic-embed-text. Invalid models return error.'
• Add error recovery guidance to tool descriptions. Example for generate: 'If model not found, call list_models to see available models. If connection fails, verify OLLAMA_HOST environment variable points to running Ollama instance (default: http://localhost:11434).'
• Add limit and offset parameters to list_models. Document: 'Returns up to 50 models by default. Use offset for pagination. Total count included in response to enable client-side pagination UI.'