opencode

The OpenCode AI agent framework architecture integrates agents, skills, tools, commands, and sessions. It provides a structured configuration system that enables users to understand operational lifecycles and debug framework processes effectively.

To create custom tools and agents in an AI framework, you need to understand the framework's extension capabilities and configuration. OpenCode provides detailed references for defining and loading agents to facilitate custom development.

Spawning a subagent fails during session management due to incorrect session lifecycle configurations or permission restrictions. Examining the framework's session lifecycle and permission system helps diagnose and resolve these subagent errors.

Yes, you can use the REST and SSE API for debugging AI agents. The framework provides REST and SSE API endpoints specifically designed to support operational debugging and real-time event streaming for agent interactions.

The best way to manage AI agent skills and configurations is through a centralized skill management system. OpenCode provides a comprehensive reference detailing how skills are defined, loaded, and integrated with agents.