Semantic Graph Tab Reference

The Semantic Graph tab builds and explores a knowledge graph of your binary, capturing function relationships, security properties, and semantic information.

Semantic Graph Tab Interface

Purpose

The Semantic Graph tab creates a rich, queryable knowledge base of your binary that:

This graph can be queried directly through the UI or used to enhance LLM responses in other tabs.

Sub-Tabs

The Semantic Graph tab contains four sub-tabs:

List View

Displays all indexed functions in a searchable table.

Column Description
Address Function start address
Name Function name (or generated symbol)
Summary LLM-generated description of function purpose
Flags Security flags (network, file I/O, crypto, etc.)
Community Detected function group/module

Click a row to: - View the function’s callers and callees - Navigate to the function in Binary Ninja

Visual Graph

Interactive visualization of function relationships.

Visual Graph

Navigation: - Pan: Click and drag - Zoom: Scroll wheel - Select: Click a node to view details

Full-text search across function names and summaries.

Enter keywords to find functions by: - Function name - Summary content - Security-related terms

Results show matching functions with relevant excerpts highlighted.

Manual Analysis

Controls for running individual analysis pipelines.

Analysis Buttons

The Manual Analysis panel provides buttons to run specific analysis types:

Button Description
ReIndex Binary Extract all functions and build call graph
Semantic Analysis Generate LLM summaries for all functions
Security Analysis Run taint analysis for vulnerability detection
Community Detection Group related functions using Label Propagation
Clear Graph Delete the semantic graph and start fresh

ReIndex Binary

Extracts function information and builds the call relationship graph:

This is the foundation for all other analyses.

Semantic Analysis

Uses the LLM to analyze each function and generate:

This runs on all functions that don’t already have summaries.

Security Analysis

Performs taint analysis to find potential vulnerabilities:

Discovered paths are stored in the graph for later review.

Community Detection

Groups related functions using the Label Propagation algorithm:

Communities help you understand the binary’s modular structure.

Graph Statistics

The bottom of the tab displays current graph statistics:

Statistic Description
Functions Number of indexed functions
Edges Number of relationships (calls, taint flows)
Communities Number of detected function groups

Using the Graph

With LLM Queries

When MCP is enabled in the Query tab, the LLM can query the semantic graph to:

For Manual Analysis

Use the List View and Search to: - Find functions matching criteria - Review security flags - Understand function purposes - Navigate the binary systematically