The server provides 20 tools with explicit schemas and descriptions visible in src/index.ts. Naming follows verb_noun conventions consistently (get_, list_, create_, update_, delete_, enable_, disable_). Descriptions are present for all tools and most parameters, though several are generic or lack implementation guidance. Schemas are properly structured with type declarations and required fields. However, significant gaps exist: (1) Output schemas are NOT documented, the response structure from each tool is undisclosed, forcing LLMs to guess at downstream field names. This breaks tool composition and forces discovery overhead. (2) Parameter descriptions are minimal and lack validation constraints (ranges, formats, patterns). (3) Error handling is mentioned for rate limits in interceptor code but not reflected in tool descriptions. (4) Some tools (enable_api, disable_api, enable_mcp, disable_mcp) lack dependency/permission documentation despite modifying system state. (5) Destructive operations (delete_server) lack confirmation or dry-run patterns. The server is functional but incomplete for production agentic use.

• Document response schemas for all 20 tools. Use comments or a schema registry to define: for get_server, return {uuid, name, description, ip, port, user, proxy_type, status, created_at}; for list_servers, return {servers: [{uuid, name, ip, status}], total_count: number, has_more: boolean}. This enables LLMs to compose tools correctly.
• Expand parameter descriptions to 50-100 characters with validation constraints. E.g., 'team_id: Numeric team identifier (string, e.g. "0" or "1"). Must exist in current workspace. Case-sensitive.' Include format, range, and error conditions.
• Implement pagination for list_teams and list_servers: add limit (1-100, default 20), offset/cursor parameters. Return total_count and next_cursor in response. Document in tool description: 'Returns up to 20 teams per page. Use offset to retrieve subsequent pages.'
• Add confirmation/dry-run support to delete_server. Implement a dry_run parameter (default true) that returns what would be deleted without actually deleting. Tool description: 'Permanently removes server and all associated resources. Set dry_run=false to confirm deletion.'
• Document permission requirements and system impact in all enable_*/disable_* tool descriptions. Add: 'Requires root-level API token. Affects all teams. Will interrupt active deployments if disabled.' Include estimated downtime.
• Expose rate-limit handling in tool descriptions. Add note: 'May return rate_limit_exceeded error if Coolify API is under heavy load. Retry after indicated seconds or wait 60s.'