Manages thin adapter boundaries and hooks to integrate AI agents with the Engram persistent memory system.
The Engram Plugin Thin skill provides a standardized architectural framework for connecting AI coding agents—including Claude, OpenCode, Gemini, and Codex—to the Engram memory core. It enforces clean separation of concerns by ensuring plugin adapters remain lightweight while delegating complex logic to the high-performance Go-based backend. This skill is essential for developers looking to implement persistent memory capture and active memory recall within their customized agent environments without introducing runtime bloat or unnecessary dependencies.
Key Features
01Multi-agent support including Claude, Gemini, and OpenCode
02Automated compatibility checks for agent-specific setup flows
03Seamless integration with Go-based memory cores via MCP and HTTP
04Dependency-free architecture for lightweight plugin scripts
051,850 GitHub stars
06Standardized adapter boundary rules for AI agent plugins
Use Cases
01Developing lightweight plugin hooks for Gemini or Codex integrations
02Wiring passive memory capture into custom Claude Code workflows
03Maintaining a shared integration contract across multiple AI coding platforms
