OpenSage Development Wiki¶

Welcome to the OpenSage Development Documentation!

OpenSage (Open Self-programming Agent Generation Engine) is an AI agent framework built on top of Google ADK.

Quick Navigation¶

• Introduction - Project overview and features
• Getting Started - Quick start guide
• Architecture - System architecture overview
• Core Concepts - Core concepts explained

Documentation Index¶

Getting Started¶

• Introduction - What is OpenSage?
• Getting Started - Environment setup and installation
• Project Structure - Project structure overview

Entry Points¶

• Entry Points - Overview of entry points
• OpenSage Web Entry - Interactive web UI workflow
• Evaluation Entry - Batch evaluation workflow
• RL Integration - RL framework integration for agent training

Core Topics¶

• Architecture - Architecture design
• Core Components - Core components explained
• Core Concepts - Core concepts (Sessions, Agents, Sandboxes)

Development Guides¶

• Development Guides - Development guide overview
• Adding Tools - How to add new tools
• Sandboxes - Sandbox backends and how to add new sandboxes
• Adding Evaluations - How to add evaluation benchmarks

Practices¶

• Best Practices - Best practices
• Common Patterns - Common code patterns
• Testing Debugging - Testing and debugging

Reference¶

• Contributing - Contributing guidelines

Key Features¶

• Session-based resource management with isolation
• Multi-sandbox support (Docker, Kubernetes)
• Security tool integration (Joern, CodeQL, GDB, Neo4j)
• Dynamic tool loading
• Evaluation framework

Getting Started¶

If you're new to OpenSage, we recommend reading in this order:

1. Introduction - Learn about the project
2. Getting Started - Set up your environment
3. Core Concepts - Understand core concepts
4. Development Guides - Start developing

Check the GitHub repository for the latest code