Four tools with basic descriptions and partial schemas. All tools have descriptions (10-200 chars), but parameter documentation is inconsistent. Input schemas present but lack comprehensive type definitions and constraints. No output schemas documented. Error handling guidance absent. Tool names follow verb_noun pattern but lack clarity on distinctions (audit_site vs validate_ucp both scan). No tool annotations (readOnlyHint/destructiveHint). Composition is reasonable (each tool has one responsibility), but chaining support is weak, no documented output fields that feed into downstream tools.

• Document output schemas for all tools. E.g., audit_site returns {has_ucp_profile: bool, ucp_capabilities: [string], ucp_version: string, profile_errors: [string], ...}. Include field types and descriptions so LLMs can chain calls.
• Rename validate_ucp to validate_ucp_profile or check_ucp_compliance to disambiguate from audit_site. Update descriptions to clarify: audit_site scores readiness gaps; validate_ucp checks schema compliance.
• Add enum constraints to parameters. E.g., caps: enum=['checkout', 'catalog', 'fulfillment', 'discount', 'order']. Replace 'comma-separated' with formal constraint. Add currency pattern: /^[A-Z]{3}$/.
• Add error handling guidance to descriptions. E.g., 'If URL is unreachable, returns error with retry guidance. If /.well-known/ucp is malformed JSON, returns profile_errors array with specific issues.'
• Add tool annotations: mark audit_site and validate_ucp with readOnlyHint=true; mark generate_profile and map_catalog with destructiveHint=true (they write files). Mark all as idempotentHint=true if re-running with same inputs produces same output.