Two tools with clear, domain-specific names (check_page, check_crawlers) and solid descriptions (145 - 160 chars, within baseline 194 avg). Input schemas present with proper types and enums for 'lang' parameter. However, output schemas are NOT documented in the source, LLMs cannot predict return structure. Parameter descriptions are minimal (e.g., 'The URL to check' lacks format guidance). No error handling guidance visible. Tool composition is sound (single responsibility), but schema completeness and error recovery patterns are missing.

• Document output schema for both tools. Specify fields (e.g., score, readiness, responseMs, findings), types (number, string, array), and whether results are paginated. Include example response structure.
• Expand parameter descriptions: 'url' should state 'Full or relative URL (http/https). Max 2048 chars. Scheme optional; defaults to https.' 'lang' already has enum constraint, good.
• Add error handling guidance to tool descriptions. E.g., 'Returns error if URL is unreachable (retryable) or malformed (user-fixable). See error.code for classification.'
• Consider migrating from STDIO to HTTP+SSE or Streamable HTTP transport to enable remote client access and hosted deployments.
• Add tool annotations (readOnlyHint: true) to both tools since they perform no mutations, helps agents reason about safety and retry logic.