prose

Execute OpenProse programs to coordinate multi-agent workflows across sessions.

Updated Apr 7, 2026
One-click install
npx skills add https://github.com/Luzzer/Kraken --skill prose-luzzer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: prose
Source: https://github.com/Luzzer/Kraken/tree/main/extensions/open-prose/skills/prose
Command: npx skills add https://github.com/Luzzer/Kraken --skill prose-luzzer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

OpenProse provides a formal, self-describing framework to orchestrate AI subagents, manage state, and coordinate complex workflows with minimal boilerplate.

Core Features & Use Cases

  • Built-in agent definitions, session spawning, imports, and parallel execution to coordinate multi-step tasks.
  • Supports multiple state backends (filesystem, in-context, SQLite, PostgreSQL) for persisting runs and enabling resumption.
  • Use cases include captain's chair orchestration, long-running workflows, and cross-project agent memory.

Quick Start

Start by drafting a .prose file that defines agents and sessions, then run with prose run to execute the workflow.

Frequently Asked Questions about prose

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

FAQPage Schema
How do I orchestrate multi-agent workflows and manage parallel sessions?

Multi-agent workflows are orchestrated by executing programs that spawn subagents, coordinating parallel or sequential tasks across sessions while managing state, persistence, and agent memory automatically.

What is OpenProse and how does it coordinate AI agents?

OpenProse is a formal framework to orchestrate AI subagents, managing state and coordinating complex workflows with minimal boilerplate through a virtual machine that handles state, persistence, and agent memory across runs.

How do I start defining agents and sessions for task orchestration?

Start by drafting a .prose file that defines agent identities and sessions using frontmatter, provide instructions in the body, then run the workflow using the command line to execute the program.

Can I use SQLite or PostgreSQL for persisting long-running workflow states?

Yes, long-running workflows support multiple state backends including filesystem, in-context, SQLite, and PostgreSQL for persisting runs and enabling resumption across sessions.

Does this orchestration framework support cross-project agent memory resumption?

Yes, the virtual machine coordinates state, persistence, and agent memory across runs, supporting cross-project agent memory and enabling long-running workflows to resume via multiple state backends.

What are the limitations of orchestrating AI subagents with a self-describing framework?

Orchestration requires drafting .prose files with frontmatter identity and body instructions, meaning workflows must adhere to a formal runtime wiring contract rather than using ad-hoc scripts.