AgentRunner (runner.ts)

@fileoverview Core conversation loop engine for open-multi-agent. {@link AgentRunner} is the heart of the framework. It handles: - Sending messages to the LLM adapter - Extracting tool-use blocks from the response - Executing tool calls in parallel via {@link ToolExecutor} - Ap

// Source: https://github.com/open-multi-agent/open-multi-agent | License: MIT | Copyright Shenzhen YuanASI Technology Co., Ltd. and open-multi-agent contributors
// Provided via justclaw.it.com; keep this notice with the code.
/**
 * @fileoverview Core conversation loop engine for open-multi-agent.
 *
 * {@link AgentRunner} is the heart of the framework. It handles:
 *  - Sending messages to the LLM adapter
 *  - Extracting tool-use blocks from the response
 *  - Executing tool calls in parallel via {@link ToolExecutor}
 *  - Appending tool results and looping back until `end_turn`
 *  - Accumulating token usage and timing data across all turns
 *
 * The loop follows a standard agentic conversation pattern:
 * one outer `while (true)` that breaks on `end_turn` or maxTurns exhaustion.
 */


... (truncated -- full source via MCP)