Server has 4 well-named tools with clear descriptions and documented schemas. All tools start with action verbs (search_, get_, add_). Descriptions are substantive (100-200 chars), explaining what the tool does and when to use it. Input schemas are present with type definitions and parameter descriptions. Tool annotations (readOnlyHint, idempotentHint) are correctly applied. However, output schemas are not formally documented, responses are JSON strings without declared structure. Error handling is minimal (no recovery guidance or categorization). Parameters lack some constraints (e.g., no enum for 'dimension', no min/max for 'limit'). The add_annotation tool correctly documents idempotency, but other tools lack explicit side-effect declarations.

• Formally document output schemas for all tools. E.g., search_documents returns {success: bool, data: {documents: [{id, title, source, body, tags, annotations}]}, meta: {result_count, duration_ms}}. Include this in tool descriptions or a separate schema block.
• Add enum constraint to 'dimension' parameter. Define valid dimensions (e.g., 'methodology', 'caveats', 'references', 'related_work') and enforce them. Update descriptions to reference the enum rather than 'e.g.'.
• Expose limit and offset/cursor parameters in search_documents and search_annotations. Allow agents to control pagination. Document the default (20) and max (e.g., 100) in descriptions.
• Enhance error responses with recovery guidance. E.g., 'Document not found (id=xyz). Try search_documents() to find the correct ID.' Categorize errors as retryable vs user-fixable.
• Expand get_document description to explain when to use it (e.g., 'Call after search_documents to retrieve full record with all annotations and tags') and what it returns.