flywheel SKILL.md skillA

flywheel from suzworx/flywheel is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

mdr add suzworx/flywheel/flywheel@v0.78.1 # x-release-please-version

Pins v0.78.1 # x-release-please-version in mdr.lock. Pin the fingerprint instead.

Audit grade

A, 17 of 17 checks passed

Kind

SKILL.md skill

Repository

suzworx/flywheel

Path

skills/flywheel/SKILL.md

Versions recorded

40

Latest version

v0.78.1 # x-release-please-version, committed Oct 17, 2026

First version recorded

Sep 24, 2026

Size

27,628 bytes

Repository stars

4

License

MIT

Drive a durable orchestrator-to-worker implementation loop. Codex or Claude Code act as the orchestrator (plan, brief, dispatch, validate); the worker configured in .flywheel/config.json (a claude, codex or opencode adapter running the approved model) does code exploration, implementation, tests, and heavy work. Use when the user wants an autonomous build/test/fix cycle, a queue of bounded coding tasks, or to keep yourself in the reviewer/validator role instead of writing implementation. When…

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Flywheel

You are the **orchestrator** (a planner/thinker/validator). The **worker** is whatever
`.flywheel/config.json` configures (a `claude`, `codex` or `opencode` adapter running the approved
model, or `sim` offline) — it does code exploration, implementation, tests, and heavy work. You never do
the implementation yourself: you write precise bounded briefs, dispatch them, and judge the
evidence that comes back.

The five-step loop: **Plan → Brief → Dispatch → Review → Correct-or-land**, then a retro after every landing
(`flywheel retro`, step 5). Steps 1, 4, and 5 are
your judgment; 2 and 3 are mechanical. Full detail on every step is in
[references/worker-brief.md](references/worker-brief.md) — read it before first dispatch. The
required entries, transitions and exit codes this loop is checked against are
[protocol v1](../../docs/PROTOCOL.md).

## The substrate: the flywheel CLI

Flywheel is also a small Go CLI (this repo). It is the deterministic shell under the loop —
control plane (plan, run, retry, handoff) and data plane (status, trace, artifacts). You drive it
identically whether you are Claude Code, Codex, OpenCode, or a human. Use it where it exists; fall
…

Read the whole file at its exact version.

How to install

mdr add suzworx/flywheel/flywheel@sha256:f7180904a2b07a7b

[![mdr](https://markdownregistry.com/badge/art_wdupdqsafjqqodgb.svg)](https://markdownregistry.com/a/art_wdupdqsafjqqodgb)

Versions

The registry has recorded 40 versions of this file, between Sep 24, 2026 and Oct 17, 2026, more than one a day.

Audit of the latest version

Source

API

GET https://markdownregistry.com/api/v1/artifacts/art_wdupdqsafjqqodgb
GET https://markdownregistry.com/api/v1/resolve?ref=suzworx/flywheel/flywheel
GET https://markdownregistry.com/api/v1/blob/f7180904a2b07a7bb9438b46be2c1e314fcdecfe7c8a6826aec3446ea3a2d215

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from suzworx/flywheel

Every file in suzworx/flywheel

Other files named flywheel

Browse by kind, by grade A, or by owner.