easymd provides 4 markdown-focused tools with basic schemas and descriptions in Chinese. All tools have input schemas with maxLength constraints and descriptions, but lack critical production-grade details: no output schemas documented, no error handling guidance, no parameter validation rules beyond maxLength, and descriptions lack WHEN/WHY context. Tool names are action verbs (extract, lint, parse, render) which is good, but descriptions are generic and don't explain dependencies or failure modes. No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles. Schemas are minimal JSON Schema with only type and description, no enums, patterns, or structured output definitions. This is typical of early-stage community servers.

• Add output schemas for each tool. E.g., extract returns {text: string, char_count: number}; parse returns {ast: object, headings: array}; render returns {html: string, warnings: array}.
• Expand descriptions to 100-200 chars with WHEN/WHY context. E.g., 'extract: Use when you need plain text for word count, summarization, or TTS. Returns stripped text without formatting.'
• Add tool annotations: mark extract/parse/render as readOnlyHint=true; mark lint as destructiveHint=true (modifies input).
• Document parameter constraints in descriptions: 'markdown: Markdown source (max 1MB). Must be valid UTF-8. Returns error if unparseable.'
• Add error handling guidance: 'If lint fails, the markdown may have syntax errors. Try parse() first to validate structure.'