[Feature] Add LangGraph framework parser for threat modeling

[lxcxjxhx]

Summary

Add support for parsing LangGraph workflow files (.py) in the threat modeling pipeline.

Problem

Agent-Wiz currently supports Autogen, CrewAI, and Swarm frameworks but lacks LangGraph support. LangGraph is widely used for building stateful, multi-actor agent applications.

Proposed Implementation

• New file: src/repello_agent_wiz/frameworks/langgraph/mapper.py
• AST-based parsing for LangGraph-specific constructs (StateGraph, nodes, edges)
• Extract agent workflows similar to existing framework parsers
• Add corresponding tests

Use Case

Security teams using LangGraph-based agent systems can automatically generate threat models without manual graph construction.

References

• LangGraph documentation: https://langchain-ai.github.io/langgraph/