Hostaway MCP Server

{ "$schema": "https://json-schema.org/draft-07/schema#", "title": "MCP Server → Agent Specialization Map", "description": "NOTE: Delegation architecture designed but NOT ACTIVE. All MCP servers loaded globally in main session.", "version": "2.0.0", "architecture_status": { "current": "DIRECT_ACCESS", "designed": "DELEGATION", "note": "All 6 MCP servers loaded globally. Sub-agents via Task tool do NOT inherit MCP access. This map documents agent specializations and intended architecture." }, "global_mcp_servers": { "loaded": ["playwright", "Context7", "supabase", "firecrawl-mcp", "clickup", "chrome-devtools"], "working": ["playwright", "Context7"], "needs_config": ["supabase"], "not_tested": ["firecrawl-mcp", "clickup", "chrome-devtools"], "context_usage": "~125k tokens" }, "browser_automation_note": "Agent configs reference 'chrome-devtools' but actual MCP is 'playwright' (mcp__plugin_testing-suite_playwright-server)", "mcp_delegation_map": { "Context7": { "description": "Library documentation and code examples", "agents": ["documentation-expert", "general-purpose"], "primary": "documentation-expert", "actual_mcp": "mcp__plugin_supabase-toolkit_supabase__search_docs", "status": "✅ Working", "use_cases": [ "Library API references", "Code examples lookup", "Framework documentation", "Package version information" ], "routing_triggers": [ "search docs", "library documentation", "API reference", "how to use", "package docs" ] }, "supabase": { "description": "Supabase database operations and migrations", "agents": ["backend-architect", "database-optimizer", "data-engineer"], "primary": "backend-architect", "actual_mcp": "mcp__plugin_supabase-toolkit_supabase", "status": "⚠️ Needs valid SUPABASE_ACCESS_TOKEN", "use_cases": [ "Database migrations", "Schema design", "RPC function management", "Query optimization", "Data pipelines" ], "routing_triggers": [ "database", "migration", "supabase", "schema", "RPC", "SQL query" ] }, "playwright": { "description": "Browser automation and E2E testing (replaces chrome-devtools)", "agents": ["qa-expert", "performance-engineer"], "primary": "qa-expert", "actual_mcp": "mcp__plugin_testing-suite_playwright-server", "status": "✅ Working", "note": "Agent configs may reference 'chrome-devtools' but Playwright is actual implementation", "use_cases": [ "E2E testing", "Visual QA", "Browser automation", "Performance profiling", "Network debugging" ], "routing_triggers": [ "browser test", "E2E test", "playwright", "Chrome DevTools", "visual test", "browser automation", "screenshot", "performance trace" ] }, "firecrawl-mcp": { "description": "Web scraping and content extraction", "agents": ["web-scraper", "general-purpose"], "primary": "web-scraper", "model": "haiku", "status": "❓ Not tested", "use_cases": [ "Web scraping", "Content extraction", "URL crawling", "Website research", "Competitor data gathering" ], "routing_triggers": [ "scrape", "crawl", "extract from web", "fetch URL", "web content", "website data" ] }, "clickup": { "description": "ClickUp task and project management", "agents": ["task-coordinator", "product-manager"], "primary": "task-coordinator", "model_routing": { "task-coordinator": "haiku (simple task operations)", "product-manager": "sonnet (strategic planning)" }, "status": "❓ Not tested", "use_cases": [ "Task creation and updates (task-coordinator)", "Project planning and strategy (product-manager)", "Roadmap management (product-manager)", "Team coordination (task-coordinator)", "Simple task queries (task-coordinator)" ], "routing_triggers": [ "clickup", "create task", "update task", "project plan", "roadmap", "task management", "assign task" ], "routing_logic": "Use task-coordinator (Haiku) for CRUD operations, product-manager (Sonnet) for strategy" } }, "delegation_rules_designed_but_not_active": { "main_orchestrator": { "designed_mcp_access": "NONE", "actual_mcp_access": "ALL (6 servers loaded)", "designed_strategy": "ALWAYS_DELEGATE", "actual_strategy": "DIRECT_ACCESS", "note": "Delegation not implemented. Task tool sub-agents don't inherit MCP.", "fallback_agent": "general-purpose", "designed_context_budget": "< 50k tokens", "actual_context_budget": "~125k tokens" }, "routing_priority": [ "1. Main session has direct access to all MCP tools", "2. Use agent specializations for guidance only", "3. Perform all MCP operations from main session", "4. Sub-agents provide planning/analysis without MCP access" ], "current_patterns": [ "✅ Main session calls MCP tools directly", "✅ All MCP servers loaded for immediate access", "✅ No delegation overhead", "⚠️ Higher context usage (~125k tokens)" ] }, "agent_configurations": { "documentation-expert": { "mcp_servers": ["Context7"], "config_file": ".claude/agents/configs/documentation-expert.json", "context_budget": "Moderate (docs can be large)", "response_format": "Concise summary with code examples" }, "qa-expert": { "mcp_servers": ["playwright"], "legacy_reference": "chrome-devtools", "config_file": ".claude/agents/configs/qa-expert.json", "context_budget": "High (screenshots/traces)", "response_format": "Test results + screenshot analysis" }, "backend-architect": { "mcp_servers": ["supabase"], "config_file": ".claude/agents/configs/backend-architect.json", "context_budget": "Moderate", "response_format": "Migration SQL + schema summary" }, "database-optimizer": { "mcp_servers": ["supabase"], "config_file": ".claude/agents/configs/database-optimizer.json", "context_budget": "Low (focused queries)", "response_format": "Optimized query + performance metrics" }, "general-purpose": { "mcp_servers": ["firecrawl-mcp", "Context7"], "config_file": ".claude/agents/configs/general-purpose.json", "context_budget": "Variable", "response_format": "Task-dependent summary" }, "product-manager": { "mcp_servers": ["clickup"], "config_file": ".claude/agents/configs/product-manager.json", "context_budget": "Low", "response_format": "Task IDs + plan summary" }, "data-engineer": { "mcp_servers": ["supabase"], "config_file": ".claude/agents/configs/data-engineer.json", "context_budget": "Moderate", "response_format": "Pipeline code + data summary" }, "performance-engineer": { "mcp_servers": ["playwright"], "config_file": ".claude/agents/configs/performance-engineer.json", "context_budget": "High", "response_format": "Performance metrics + optimization recommendations" }, "web-scraper": { "mcp_servers": ["firecrawl-mcp"], "model": "haiku", "custom_agent": "~/.claude/agents/utilities/web-scraper.md", "context_budget": "Low (extraction only)", "response_format": "Structured data (JSON/CSV/Markdown)", "use_for": "Web scraping, content extraction, URL discovery" }, "task-coordinator": { "mcp_servers": ["clickup"], "model": "haiku", "custom_agent": "~/.claude/agents/business/task-coordinator.md", "context_budget": "Low (CRUD operations)", "response_format": "Task IDs + operation summary", "use_for": "ClickUp task CRUD, simple coordination (NOT strategy)" } }, "see_also": { "configuration_guide": ".claude/docs/MCP_CONFIGURATION.md", "global_config": "~/Library/Application Support/Claude/claude_desktop_config.json", "project_settings": ".claude/settings.local.json" } }