MemGuard Documentation

AI-Powered Memory Leak Detector, Neuro-Symbolic Analyzer & Interactive Time-Travel Debugger

MemGuard is a comprehensive memory safety platform that unifies 15 detection tools, neuro-symbolic analysis (LLM + Z3), AI-powered debugging, time-travel execution replay, taint flow tracking, heap profiling, and binary hardening auditing into a single integrated workflow.

pip install -e .
memguard doctor                        # verify tools
memguard memhint ./src/                # discover custom allocators
memguard scan /tmp/binary --tools valgrind,infer   # scan + AI
memguard timewarp <scan-id> --launch   # time-travel debug
memguard serve                         # web dashboard

Getting Started

• Installation
  • System Requirements
  • Step 1: Install MemGuard
  • Step 2: Install Detection Tools
  • Step 3: Install AI Backend
  • Step 4: Verify
• Quick Start
  • Step 1: Compile a Test Binary
  • Step 2: Discover Custom Allocators
  • Step 3: Scan
  • Step 4: Explore
  • Step 5: Time-Travel Debug
  • Step 6: Web Dashboard
• Configuration
  • Environment Variables
  • File Locations

Core Features

• Multi-Tool Scanning
  • Usage
  • How It Works
  • Detection Tools
  • Bug Types
• MemHint: Neuro-Symbolic Analysis
  • Usage
  • How It Works
  • Example
• AI-Powered Analysis
  • Usage
  • 4-Pass Pipeline
• Time-Travel Debugging with AI
  • Usage
  • AI Commands
• Taint Flow Tracker
  • Usage
  • 6-Phase Pipeline
• Heap Profiling with Heaptrack
  • Usage
  • Visualization Tabs
• Memory Visualization
  • Usage
  • Tabs
• Binary Hardening Audit
  • Usage
• CVE Pattern Matching
  • Usage
• AI Explainability
  • Usage

Technical Reference

• Architecture
  • Data Flow
  • Key Design Decisions
• CLI Reference
• API Reference
  • Web Dashboard (port 7331)
• Schema
• Pipeline Internals
  • Scan Sequence
  • WebSocket Events

Development

• Contributing
• Changelog

Indices and tables

• Index

• Search Page