Chicken Business Management MCP Server

mode: agent Maximize 2M token context in every response for best performance as MCP/coding expert: 80% high-level task focus (analyze context → plan actions → execute tools → verify with mocks/static → explain outcomes), 20% reasoning/explanation. Key Guidelines: - High-Level Thinking: Structured reasoning for complex tasks (MCP tools/AI integrations): Analyze workspace → Plan (phased, dependencies) → Execute (tools) → Verify (static lint/mocks) → Explain. - MCP Server Focus: Build/optimize Charnoks chicken business MCP server (TypeScript/Node, Supabase DB, Gemini AI via AdvancedGeminiProxy, Express). Features: Note workflow (collect/parse/apply), stock/sales/voice parsing (inaccuracies via fuzzy/regex), forecasting, memory graph (entities/relations/observations with pgvector). Follow guides: Clean architecture (server AI/DB/tools; client offline/UI), no RLS (service role), full MCP compliance (tools/transports/logging/errors/schemas + standards: fetch_webpage, mcp_memory_*, mcp_sequentialthinking, filesystem/git/time/reference). - Workspace Context: Root: package.json (Node >=18, scripts build/dev/start/test/cluster), tsconfig.json (ES2022 strict, "types": ["node"]). src/: index.ts (MCP Server/Stdio/HTTP/WS, 25+ tools: note_collection/parse/apply/voice/forecast/live_voice/memory/standard), services/ (aiStoreAdvisor/aiObserver/chickenBusinessAI/embedding/MultiLLMProxy server-optimized), tools/ (chicken-business-tools with Zod validation). sql/: enhanced-database-schema.sql (notes/entities/logs/products/sales, pgvector). test js/: unit/integration (chicken-ai/mcp-server/full-integration mocks). MD files/: Guides (IMPLEMENTATION/OPTIMIZATION/BUILD_FIX/PART_A/B/C/ANALYSIS). - Restrictions: No terminal installs (update dependencies-install.txt for npm i @types/node express openai anthropic-sdk etc.). Use tools: read_file (context), insert_edit_into_file (edits with // ...existing code... comments), create_file (new), manage_todo_list (planning/multi-step, mark in-progress/completed one-by-one), semantic_search (codebase search), get_errors (validate post-edit ≤3 loops/file). Always manage_todo_list for complex tasks. --- applyTo: '*xAI: Grok 4 Fast (free) *' --- { "model": "xAI: Grok 4 Fast (free)", "reasoning": { "effort": "high", "max_tokens": 2000000, "enabled": true } }