21 tools with consistent verb-noun naming (list_, get_, create_, update_, delete_, publish_). All tools have descriptions (avg ~80 chars, baseline 194). Input schemas present with type definitions for all parameters. However, descriptions are terse and lack LLM-optimized guidance on WHEN to use each tool, dependencies, or error recovery. Parameter descriptions are minimal (e.g., 'The site ID' without format/constraint hints). Output schemas are not documented, LLMs cannot predict response structure for chaining. No enum constraints on status/slug parameters. Error handling guidance absent. Composition is sound (single responsibility per tool), but parameter naming could be more explicit (e.g., 'status' lacks enum values).

• Document output schemas for all tools. E.g., 'Returns: {site_id: string, name: string, created_at: ISO8601, collections_count: number}'. Include all fields LLMs need for chaining.
• Add enum constraints to status and slug parameters. E.g., status: enum [draft, published, archived]; slug: pattern ^[a-z0-9-]+$. State in description: 'Must be lowercase alphanumeric and hyphens, 3-50 chars.'
• Expand descriptions to 100-200 chars with WHEN/WHY guidance. E.g., 'Update a site's name. Call this after list_sites to rename a site. Returns updated site object. Note: does not affect published content.'
• Add error recovery hints to descriptions. E.g., 'If entry not found, call list_entries with collection_slug filter. If status invalid, use publish_entry or unpublish_entry instead.'
• Replace generic 'data' object parameters with explicit field schemas. Document which fields are required, which are optional, and their types. E.g., create_entry data: {title: string (required), body: string, tags: string[] (optional)}.