agentsclimarketplace

Setup working genius

Skill donald-ada/workinggenius/skills/setup-working-genius

A Claude Code development workflow plugin: walks every piece of work through the six Working Geniuses (Wonder, Invention, Discernment, Galvanizing, Enablement, Tenacity) so no stage gets skipped.

Install
npx -y skills add donald-ada/workinggenius --skill setup-working-genius

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Optional per-repo configuration — pin the work-file directory and verify commands so every stage runs them the same way.

SKILL.md

2.9 KB, as published. Nobody here has run it

Setup Working Genius

The workflow runs fine on defaults (.genius/ for work files, verify commands discovered per run). This skill pins those choices per repo so no session has to rediscover them. Run it once, or never.

Run it as a conversation, not a script: look first, propose what you found, and let the user correct each choice before anything gets written.

Process

1. Explore

  • CLAUDE.md / AGENTS.md at the repo root — which exists? Is there already a ## Working Genius section?
  • .genius/ — any work files already in flight?
  • The project's task runner (package.json scripts, Makefile, justfile, pyproject.toml, …) — find the likely typecheck, test, and lint commands.

2. Ask, one decision at a time

A — Work-file directory. Default .genius/ at the repo root. Also ask: commit work files or gitignore them? Recommend committing — done files are decision history, and gates only bind what's visible to every session.

B — Verify commands. Propose the discovered typecheck / test / lint commands; let the user correct them. These are what Enablement runs each cycle and Tenacity runs fresh at close-out.

C — Gate enforcement. How hard should the Stop hook hold the gate rule when a stage ran past an unchecked, unskipped gate? warn (default): blocks ending the session once per distinct bypass state — the repair message lands, then the user can still stop. block: every stop attempt re-blocks until the bypass is repaired or recorded as skipped. Recommend warn for solo exploratory repos and block where the work files are shared history other sessions build on. Only block needs writing — warn is what no line means.

3. Write

Pick the file: CLAUDE.md if it exists, else AGENTS.md if it exists; if neither, ask which to create — never create one when the other already exists. If a ## Working Genius section already exists, update it in place — don't append a duplicate, and preserve any Lessons: list Tenacity has grown there (it's project memory, not setup output).

## Working Genius

Work files: `.genius/` (committed). Flow: /wonder → /invent → /discern → /galvanize → /enable → /tenacity; type /genius for status.
Gate enforcement: block

Verify commands:
- typecheck: `<command>`
- test: `<command>`
- lint: `<command>`

Don't seed a Lessons: list — Tenacity creates it lazily, the first time a lesson recurs across work and earns promotion.

If the user chose gitignore, add the directory to .gitignore.

4. Done

Tell the user setup is complete, and that editing the ## Working Genius section directly is the normal way to change these later — re-running this skill is only for starting over.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.