Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and output structure. Tool names follow verb_noun pattern (resolve-*, get-*), which is good. However, descriptions lack actionable guidance on prerequisites and error recovery. Parameter descriptions are present but sparse. No documented output schemas. The 'tokens' parameter has a confusing minimum constraint (5000) that contradicts the description's 'default: 5000'. Error responses are generic text blobs rather than structured guidance. No evidence of pagination, batching, or composition patterns.

• Document output schemas for both tools. resolve-library-id should return a structured list of {id, name, description, state} objects. get-library-docs should return {content, tokens_used, topic_matched, available_topics}. This enables LLMs to extract and chain data.
• Fix the 'tokens' parameter constraint. If the default is 5000, either remove the min(5000) constraint or change the description to 'Minimum 5000 tokens; default 5000'. Clarify the actual range (e.g., 5000 - 100000).
• Add structured error responses with error codes and recovery hints. E.g., {error: 'LIBRARY_NOT_FOUND', message: '...', suggestion: 'Call resolve-library-id to find valid IDs'}. Implement pattern:recovery-guide.
• Clarify whether 'libraryName' is required or optional. If optional, document the default behavior (e.g., 'Returns all finalized libraries sorted by popularity'). If required, mark it as such in the schema.
• Add pagination to resolve-library-id. Accept 'limit' (default 20, max 100) and 'offset' parameters. Return {libraries: [...], total: N, has_more: bool}. This prevents context window exhaustion.