GhidrAssist integrates LLMs into Ghidra to provide intelligent reverse engineering assistance. It supports any OpenAI v1-compatible API including local models (Ollama, LM-Studio) and cloud providers (Anthropic, Azure). Core features include code explanation with security analysis, multi-turn conversational queries, and direct LLM queries with context from current code location. The system includes a Graph-RAG knowledge system with hierarchical semantic representation of binary analysis (Statement through Binary levels), community detection via Leiden algorithm for module discovery, and comprehensive security feature extraction for network, file I/O, crypto APIs and string patterns.