OpenProse provides a contract-driven runtime for multi-step agent workflows. This skill lets the agent interpret prose commands, load contract-markdown or ProseScript files, select a state backend, validate and compile manifests, and execute or lint services and systems without shelling out to external binaries. Typical uses include running .prose.md services, compiling manifests, linting contracts, and inspecting run state.
Use OpenProse when a task requires reusable multi-agent orchestration, has three or more distinct steps, or needs an auditable contract and persistent run state. Activate when the user invokes prose commands, references .prose.md files, or asks to run/lint/inspect/upgrade Prose sources. Decline for one-shot Q&A.
Likely hosts and tooling: Claude Code, Codex, Copilot-style tooling, and other agent runtimes that support subagent spawning or command routing.
OpenProse is a multi-agent orchestration framework using contract-markdown (.prose.md) files to define responsibilities, functions, and wired DAGs. The SKILL.md is exceptionally detailed with clear command routing, format detection, host primitive adapters, and progressive disclosure via referenced docs. No bundled scripts to test. The skill promotes a disciplined approach to agent workflows but requires OpenProse tooling (reactor binary, deps) for full functionality.
Clean skill with no security concerns. Very thorough documentation and well-structured architecture. The 'giving back' section is notable — thoughtfully designed to prevent spam/abuse. No scripts present for runtime testing. The skill is substantial but niche — primarily useful for teams already invested in the OpenProse paradigm.