Prime Agent Documentation (prime-agent-docs)

Overview

Prime Agent is an RLM-native (Recursive Language Model) coding and research harness built around a persistent IPython kernel, recursive subagents, durable sessions, and a multi-process local runtime.

Core Architecture

• Persistent Kernel: Executes code interactively in a stateful IPython environment across turns.
• Recursive Subagents: Spawns isolated or shared subagent sessions to solve subtasks autonomously.
• Daemon & Runtime: Decouples UI clients from background execution workers with IPC and ZeroMQ transports.
• Durable Sessions: JSONL-backed tree and timeline storage with branching, compaction, and replayability.
• Flexible Integration: Supports interactive TUI, headless CLI, SDK (Node.js/TypeScript), ACP (Agent Client Protocol), and RPC (stdin/stdout JSONL).

When to Use This Skill