Multi Agent Runtime — Development skill for Claude Code

Multi-Agent Runtime

[English](./README.md) | [简体中文](./README.zh-CN.md)

A small, extraction-friendly runtime for running role-based multi-agent workspaces on top of the official Claude Agent SDK and Codex SDK.

This package gives us one unified protocol for:

• defining role agents and coordinator roles
• accepting a workspace-level user turn and routing it across the team
• modeling workspace members, public activity, and claim policies
• declaring template-level workflows with stages, nodes, edges, artifacts, and completion rules
• observing task lifecycle as events
• validating generated artifacts in end-to-end tests

The repository now contains:

• a TypeScript reference implementation
• a Rust protocol/core implementation intended for embedding into Cteno

The current adapters are:

• ClaudeAgentWorkspace for Anthropic Claude Agent SDK
• CodexSdkWorkspace for OpenAI Codex SDK

The protocol is intentionally generic enough to support a future Cteno-native adapter.

What It Does

`@cteno/multi-agent-runtime` treats a workspace as:

• one coordinator role
• multiple named workspace members
• explicit role dispatches such as prd, finance, or scout
• optional claim-based assignment rules
• a unified event stream for workspace, member, activity, dispatch, message, tool.progress, and result

That maps well onto Claude's session-centric model while keeping our authoring model role-centric.

Included Templates

`Coding Studio`

A software delivery workspace.

Roles:

• pm
• prd
• architect
• coder
• tester
• reviewer

Typical outcome:

• PRDs
• implementation notes
• code changes
• test reports

`OPC Solo Company`

A one-person company staffed by specialist digital operators.

Roles:

• ceo
• finance
• tax
• admin
• recruiter

Typical outcome:

• operating checklists
• finance docs
• tax prep handoff notes
• admin SOPs

`Autoresearch`

A research-oriented workspace for scouting, synthesis, and lightweight experiment