POC: Agentic System using LangGraph

[smokeyScraper]

🌟 Feature Description

Create a Proof of Concept (POC) for the Devr.AI agentic system using LangGraph framework, implementing minimal versions of DevRel Agent and GitHub Agent with basic inter-agent communication.

🔍 Problem Statement

Need to evaluate LangGraph's capabilities for:

• Building stateful agents with conversation memory
• Tool integration (web search, database queries)
• Inter-agent communication between DevRel and GitHub agents
• Handling Discord commands and GitHub events
• Managing agent state across multiple interactions
• Handling general user FAQs

🎯 Expected Outcome

A working POC demonstrating:

• DevRel Agent responding to Discord commands (!verify, !create-issue, !help-issue) along with access to read each and every message
• GitHub Agent analyzing issues and PRs
• Basic RAG functionality with vector database (can be initially tested with test data)
• Agent-to-agent communication flow with proper graph flow
• State persistence across conversations

📷 Design Architecture

A workaround for something minimal as shown in the image.

Success Metrics:

• Minimalist POC to evaluate LangGraph for the workflow with proper handling of user-specific as well FAQ-related queries.

---

🚨 Full Implementation to be extended later

---