run.prose.md

OpenProse is an open-source language and framework for writing multi-agent workflows as declarative contracts in Markdown. A contract states the work, supplies context, limits scope, and composes reusable contracts through ordinary Markdown links; contracts can also take other contracts as parameters to define relationships such as execution and independent review. An agent interprets the selected contract using its existing harness and tools, producing results and fulfillment evidence, while an evaluator can assess which requirements are satisfied, unmet, or unresolved. The workflows can be composed and triggered by events such as webhooks, pull requests, or cron jobs, and can run with harnesses including Claude Code and Codex. The kernel and versioned libraries are public Markdown, the project is described as experimental, and it can run locally via npx or through the hosted run.prose.md application.

OpenResearch is a local-first workspace for research agents and autoresearch, available as a desktop application and a macOS/Linux CLI. It turns Claude Code, Codex, or OpenCode into agents that can review literature, develop hypotheses, run experiments, and produce research artifacts. The workspace assigns independent agent sessions and isolated Git worktrees to parallel research directions. It tracks experiment variants in a Git-native experiment tree, archives each run against its recorded commit, and keeps logs, diffs, files, results, and artifacts tied to the work that produced them. Its autoresearch loop can propose an idea, modify code, launch an experiment, inspect evidence, and choose the next direction. The same committed source snapshot can run locally or through SSH, Slurm, Kubernetes, Ray, Hugging Face Jobs, Modal, Tinker, or managed OpenResearch compute. By default, projects and run data remain on the user's machine in a local SQLite store, with a browser dashboard served on localhost. An OpenResearch account is used for service-owned capabilities such as organizations and managed compute. The CLI also installs an OpenResearch skill into supported coding agents and provides commands for projects, runs, logs, experiments, discovery, and paper lookup.

Open Pstack is an unofficial, plugin-based workflow kit that brings Lauren Tan's pstack to Claude Code and Codex. It gives coding agents engineering rules, task-specific workflows, focused skills, and small local tools rather than providing a new model or hosted service. Its main poteto-mode workflow reads a task, selects an appropriate workflow, learns how the existing system works, compares designs when needed, favors small changes, and can use multiple models to challenge important decisions. It runs the code and checks behavior as a user would instead of stopping at passing tests, then can continue through review and continuous integration to prepare a pull request. Additional skills cover system explanation, architectural decisions, competing implementations, design interrogation, verification-skill creation and maintenance, pull-request supervision, and reflection. The repository supports installation as a Claude Code or Codex plugin and shares the same skill set between those applications. It is distributed under the MIT license and tracks the upstream pstack project while adapting its skills for Claude Code and Codex.

Munder Difflin is a free, open-source desktop multi-agent harness that runs multiple terminal-based AI coding agents as coordinated local agents. It wraps agents including Claude Code, OpenAI Codex, Gemini CLI, Qwen, OpenCode, GitHub Copilot CLI, and custom commands, working with users' existing subscriptions and hourly usage limits. Each agent runs as a real node-pty process and is rendered through xterm.js, while an Electron and React interface displays sessions as avatars on a Pixi.js office floor. A GOD orchestrator called Michael assigns and routes work, adjudicates agent messages, and escalates spending, destructive operations, scope changes, and other configured approvals. Agents coordinate through a local Git-backed hive of plain-file memory, atomic mailboxes, a shared blackboard, and an append-only event log; a markdown-first semantic memory layer supports recall across sessions. Optional Git worktrees isolate parallel agents.