open-prose

Orchestrate multi-agent AI workflows with a declarative VM.

3|Updated Jan 20, 2026
One-click install
npx skills add https://github.com/agusmdev/burntop --skill open-prose
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: open-prose
Source: https://github.com/agusmdev/burntop/tree/main/.claude/skills/open-prose
Command: npx skills add https://github.com/agusmdev/burntop --skill open-prose

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

OpenProse provides a declarative framework to orchestrate multi-agent AI workflows, turning English prompts into structured OpenProse programs and executing them via a VM, enabling complex automation across tools.

Core Features & Use Cases

  • Agent templates and sessions: Define reusable agents and spawn sessions to perform tasks.
  • Control flow constructs: Parallel, loops, conditionals, blocks, and pipelines to coordinate tasks.
  • Context propagation: Bind outputs from one session to inputs of others with explicit context.
  • Execution via Task tool: Each session runs as a real subagent, enabling end-to-end experiments and automation.
  • Use cases: Build data processing pipelines, code reviews, experiments, and research automation across tools.

Quick Start

Create a tiny OpenProse snippet that defines an agent and runs a single session to illustrate the workflow. For example: session: researcher prompt: "Analyze the repository"

Frequently Asked Questions about open-prose

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I orchestrate multi-agent AI workflows with parallel and looped tasks?

Multi-agent AI workflows are orchestrated using a declarative VM that runs programs with control flow constructs like parallel execution, loops, and pipelines. This enables coordinating complex sessions across various tools.

What is the best way to define reusable agents for AI session orchestration?

Reusable agents for AI session orchestration are defined using declarative agent templates. You create an agent definition and spawn sessions to execute specific tasks, binding outputs from one session to inputs of others via explicit context propagation.

Can I use the Task tool to execute AI agent sessions end-to-end?

Yes, the Task tool executes each AI agent session as a real subagent. This allows the declarative VM to run multi-agent programs end-to-end, enabling complex automation and experiments across tools.

How do I coordinate parallel tasks and conditionals in an AI automation pipeline?

Parallel tasks and conditionals in an AI automation pipeline are coordinated using declarative control flow constructs. The VM supports blocks, pipelines, and loops to sequence sessions and manage task execution across tools.

Do I need a specific framework to run data processing pipelines with multiple AI agents?

You need a declarative framework to run data processing pipelines with multiple AI agents. This approach turns English prompts into structured programs executed by a VM, satisfying requirements for structured execution models without external dependencies.

What are the limitations of using a declarative VM for AI session orchestration?

A declarative VM for AI session orchestration limits execution to its supported control flow constructs like parallel, loops, and conditionals. Complex automations must fit within these defined blocks, pipelines, and session templates.