Smart Routing Implementation - Minimal Agent Optimization - 150,000 GURU

[evahteev]

Description

Implement smart routing that separates tool responsibilities between a minimal default agent (Q&A, KB search) and a specialized bot assistant sub-agent (bot control, menus, settings). This is a 20% effort → 80% value optimization that removes tool pollution from the default agent.

Key Features:

• Minimal default agent (KB search + support tools only)
• Specialized bot assistant agent (full tool access)
• Context-aware routing logic
• Reduced agent confusion and improved response quality

Impact:

• Reduces tool count from 12-15 tools → 2-4 tools for default agent
• Eliminates menu/workflow/twitter tool confusion
• Agent focuses on Q&A and KB search
• Improved response accuracy

Reward

🎉 150,000 GURU tokens

Getting Started

1. Review implementation plan: bot/docs/features/smart-routing-implementation-plan.md
2. Check agent factory: luka_bot/agents/agent_factory.py
3. Review tool definitions: luka_bot/tools/
4. Understand current agent configuration

Implementation Phases

Phase 1: Minimal Agent (75,000 GURU) ⭐

START HERE

• Remove menu, workflow, Twitter tools from default agent
• Update static_tools list in agent_factory.py
• Test with Q&A and KB search queries

Phase 2: Bot Assistant Agent (75,000 GURU)

• Create specialized create_bot_assistant_agent() function
• Implement routing logic based on intent
• Add context-aware tool selection

Acceptance Criteria

• Default agent has only KB search + support tools
• Bot assistant agent has full tool access
• Routing logic correctly selects agent
• No regression in existing functionality
• Improved response quality for Q&A queries