This MCP server exposes 8 tools for web search, content extraction, and note management. Tool definitions are present with schemas and descriptions, but quality is inconsistent. Naming follows verb_noun conventions (tavily_search, serper_scrape, note_reader, note_editor). Descriptions range from adequate (serper_search: 219 chars) to comprehensive (note_editor: 598 chars with streaming guidance). However, parameter descriptions are present but sometimes generic, and output schemas are not documented. Error handling is minimal, no recovery guidance or categorization visible in the tool definitions. The Chinese-language descriptions are clear but suggest a regional-specific tool design. Most concerning: no visible input validation, no pagination documentation for potentially large result sets (e.g., tavily_search, serper_search), and note_editor accepts many mutually exclusive parameters without explicit documentation of dependencies.

• Document output schemas for all tools. For search tools (tavily_search, serper_search, serper_news), document the structure of each result object (title, url, snippet, etc.) and any metadata (total_count, next_cursor, etc.).
• For note_reader, document the returned fields (id, title, content, created_at, updated_at, metadata) and whether content is returned as plain text, markdown, or other format.
• For note_editor, document the response after editing (success flag, updated_note object, preview_content if save_immediately=false).
• Add pagination support documentation. Specify whether search tools support offset/limit, page/per_page, or cursor-based pagination. Document max_results behavior (is it a hard cap, or can the LLM request more?). Return total_count or next_cursor in responses.
• Explicitly document parameter interdependencies in note_editor. Create a table or structured text in the description: 'edit_type=replace requires content. edit_type=replace_lines requires content, start_line, and optionally end_line. edit_type=replace_text requires content, search_text, and replace_text.' Repeat this in each affected parameter description.
• Add error handling guidance to all tool descriptions. E.g., 'Returns 404 if URL is invalid or unreachable. If this happens, try a different URL or use serper_search to find related links.' For note_reader/editor, document: 'Returns 404 if note_id not found. Use search_title to find the note first.'
• Clarify note_editor's 'save_immediately' parameter. Explain: 'false=preview mode (returns proposed changes without saving); true=save to database immediately. Use false to show the user the changes before committing.'

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry