Three tools with complete input schemas and clear descriptions. All tools follow verb_noun naming (search_images, get_image, get_tags). Descriptions are adequate (100-180 chars) and explain what each tool does. However, output schemas are not documented, LLMs cannot see what fields to expect from responses. Parameter descriptions are present but lack constraints (e.g., no min/max for limit, no enum for namespace). Error handling is minimal, no recovery guidance or actionable error messages. No tool annotations (readOnlyHint, idempotentHint) despite all being read-only operations.

• Document output schemas for all three tools. Add a 'returns' section to each tool definition showing the structure: search_images returns {total: number, images: [{name, description, stars, pulls, is_official, is_automated}]}. This enables LLMs to extract fields correctly.
• Add min/max constraints to numeric parameters. Update limit schema: {type: 'number', minimum: 1, maximum: 100, description: 'Number of results to return (default 10, max 100)'}. Add pattern or enum to namespace: {type: 'string', enum: ['library', 'bitnami', ...] or pattern: '^[a-z0-9_-]+$'}.
• Add tool annotations to the tool definitions. Include readOnlyHint: true and idempotentHint: true for all three tools so agents know they are safe to retry and do not modify state.
• Enhance error messages with recovery guidance. When get_image fails with 404, return: 'Image not found: {namespace}/{name}. Try search_images(query="{name}") to find the correct repository.' This guides the agent to the next step.
• Add examples to parameter descriptions showing valid inputs. E.g., namespace: 'Repository namespace (e.g., "library" for official images, "bitnami" for Bitnami images)' and name: 'Repository name (e.g., "nginx", "redis", "postgres")'. Avoid listing these as the only valid options; use them to illustrate the format.