BillManager MultiAgent

Chen Mingwei, 26, 'connector' between large models and IoT. Focused on landing cutting-edge AI algorithms into the real physical world, making models evolve in real-time on edge devices. Active in open source community, likes building reusable, playable modules.

A multi-agent collaborative bookkeeping system built on OpenAgents framework, achieving income/expense record querying, adding, modifying and deleting through natural language conversation. Innovatively designed 'intent understanding - database operation' separation: Assistant Agent focuses on conversation and intent recognition, MySQL Expert Agent focuses on SQL generation and execution, collaborating through event-driven mechanisms for professional division and security isolation.

Planning two Agents collaborating through OpenAgents event system and Mods: Assistant Agent as user-facing frontend handles natural language requests (like 'last week expenses', 'received 100 yesterday'), recognizes operation intent (query/add/modify/delete) and date range, forwards to MySQL Expert via db.operation events. MySQL Expert Agent as database expert responds to Assistant's event requests, generates secure SQL, executes operations, returns results via db.result events. Core features: Event-driven collaboration via send_event() for async decoupled communication, workspace.messaging Mod for collaboration foundation, natural language date parsing, SQL security validation.

① Assistant Agent: Handles user conversation, bookkeeping intent recognition (query/add/modify/delete), date range extraction, delegating to MySQL Expert via events. ② MySQL Expert Agent: Receives database operation requests, generates secure SQL, executes operations, returns results via events. ③ Two Agents collaborate through OpenAgents event system for complete async decoupled flow. ④ Custom mysql_tools.py encapsulating database connection, SQL execution, security validation and date parsing.