UWAS exposes 12 tools with basic schemas and descriptions, but falls short of production quality. All tools have descriptions (10-80 chars, mostly under 50), which is below the 194-char baseline for A+ tools. Input schemas are present but minimal: most tools have empty properties (no parameters), and the one parameterized tool (cache_purge, domain_get) lacks detailed constraints. No output schemas are documented, LLMs cannot predict response structure for chaining. No error handling guidance, no parameter validation rules, no recovery hints. Tool names are verb-noun (good), but descriptions are terse and lack WHEN/WHY context. No tool annotations (readOnlyHint/destructiveHint) despite clear risk classifications in metadata. Overall: functional but well below production baseline.

• Add output schemas to all tools. Document the structure of domain_list response (array of domain objects with fields: host, type, ssl_mode, etc.), stats response (object with keys: requests, cache_hits, connections, etc.). This enables LLM chaining.
• Expand descriptions to 100-200 chars. Example: 'domain_list' → 'List all configured domains with their type (static, php, proxy, app, redirect) and SSL mode (disabled, letsencrypt, custom). Use this to discover available domains before calling domain_get for details.'
• Add tool annotations. Mark cache_purge with destructiveHint=true and idempotentHint=true (purging the same tag twice is safe). Mark all read-only tools with readOnlyHint=true.
• Document parameter constraints. cache_purge.tag: 'Tag to purge (alphanumeric, max 255 chars; empty string purges all). Idempotent, safe to retry.' domain_get.host: 'Domain hostname (FQDN format, e.g. example.com; case-insensitive).'
• Add error recovery guidance. E.g., cache_purge: 'If tag not found, returns success (idempotent). If cache is locked, returns 503 Unavailable, retry after 5 seconds.' domain_get: 'If domain not found, returns 404 with available domains in error message.'