-
Notifications
You must be signed in to change notification settings - Fork 4
Open
Labels
agenticAgentic AI featuresAgentic AI featuresepicLarge feature spanning multiple user storiesLarge feature spanning multiple user storiespriority:lowLow priority - when time permitsLow priority - when time permits
Description
Overview
This epic consolidates all Agentic AI features deferred to Phase 4 (Q2-Q4 2025). The design phase is complete; implementation awaits core RAG quality improvements.
Design Documents
docs/design/agent-mcp-architecture.md- Core agent architecturedocs/design/mcp-context-forge-integration.md- MCP integration strategy
Consolidated Issues (Closed → Tracked Here)
The following issues were closed and consolidated into this epic:
Agent Framework (Closed)
- Implement Base Agent Framework for Agentic AI #121 - Implement Base Agent Framework ❌ Superseded by design docs
- Build Tool Integration Framework for Agents #122 - Build Tool Integration Framework ❌ MCP replaces custom tools
- Implement Specialized Agent Types #123 - Implement Specialized Agent Types ❌ Implementation detail
- Implement Advanced Agent Capabilities - Self-Reflection and Learning #126 - Advanced Agent Capabilities ❌ Deferred to Phase 4
Agent Epics (Closed)
- EPIC-001: Agent Orchestration Framework #192 - EPIC-001: Agent Orchestration Framework ❌ Replaced by consolidated design
- EPIC-002: Workflow Engine #193 - EPIC-002: Workflow Engine ❌ Deferred to Phase 4
Still Open (Track Here)
- EPIC-003: Memory & Context Management #194 - EPIC-003: Memory & Context Management
- EPIC-004: Chain-of-Thought Reasoning #195 - EPIC-004: Chain-of-Thought Reasoning (Note: CoT for search is done)
- EPIC-005: Knowledge Graph Integration #196 - EPIC-005: Knowledge Graph Integration
- Epic: Podcast Generation and AI Evaluation Feature #240 - Epic: Podcast Generation and AI Evaluation
- Enhancement: Agentic Podcast Generation with AI Agents #241 - Enhancement: Agentic Podcast Generation
- 🤖 Feature: Agent Discovery & Orchestration Interface #244 - Agent Discovery & Orchestration Interface
- 🏗️ Architecture Decision: Agent Orchestration Framework Selection & Integration Strategy #245 - Architecture Decision: Agent Framework Selection
Prerequisites
Before starting Phase 4:
- Core RAG quality issues resolved (Epic [EPIC] RAG Quality & Search Accuracy Improvement #686)
- Backend performance optimized (Epic [EPIC] Backend Performance Optimization #687)
- Production deployment stable
Phase 4 Scope
4.1 Agent Framework Foundation
- Base agent interface implementation
- MCP tool integration layer
- Agent registry and discovery
4.2 Workflow Engine
- Task orchestration
- Multi-step workflows
- Error handling and recovery
4.3 Memory & Context
- Agent-specific context management
- Cross-session memory persistence
- Context window optimization
4.4 Knowledge Graph
- Entity extraction pipeline
- Graph storage integration
- Semantic reasoning layer
4.5 Advanced Capabilities
- Agent self-reflection
- Learning from feedback
- Multi-agent coordination
Timeline
- Design: Complete (Nov 2025)
- Implementation: Q2-Q4 2025 (after Phase 2-3)
- Dependencies: Epics [EPIC] RAG Quality & Search Accuracy Improvement #686, [EPIC] Backend Performance Optimization #687, [EPIC] Conversation System Refactoring #688
Success Criteria
- Agents can execute multi-step tasks
- MCP tools integrated
- Memory persists across sessions
- Knowledge graph queryable
- Self-reflection improves responses
Metadata
Metadata
Assignees
Labels
agenticAgentic AI featuresAgentic AI featuresepicLarge feature spanning multiple user storiesLarge feature spanning multiple user storiespriority:lowLow priority - when time permitsLow priority - when time permits