agentleFS
jocarrdjocarrd/crux-agents/skills/wizard/SKILL.md

wizard

Skill4 starsChanged 18 days agoMITView on GitHub

Generate an interactive bash wizard that walks a person, step by step, through a manual procedure the agent cannot do: provisioning infrastructure, pulling credentials from someone else's dashboard, putting secrets into CI, or running a one-off migration. It opens each URL, says what to copy, collects the values and writes them where they belong. Use when a third-party service has to be configured, when a deployment depends on secrets someone has to go and fetch, when an irreversible infrastructure change is coming, or when the user says "no me acuerdo de los pasos" or that they already explained this last time.

  • Reads credentials

170 other projects use a version of this, 137 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use setup-ts-deep-modules, writing-beats and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • liufeng1976Skill

    setup-ts-deep-modules

    Wire dependency-cruiser into a TypeScript repo so each package is a deep module — implementation hidden in subfolders, reachable only through its entry-point files. User-invoked.

    Also in 42 projects that use this

  • LesaniSkill

    writing-beats

    Shape an article as a journey of beats, choose-your-own-adventure style. The user picks a starting beat from the raw material, you write only that beat, then offer options for where to pivot next, beat by beat, until the article reaches a natural end. Use when the user has raw material and wants to assemble it as a narrative rather than an argument.

    Also in 56 projects that use this

  • RoentekSkill

    loop-me

    Grill me about specs for the workflows I want to build, within this workspace.

    Also in 49 projects that use this

  • ryotaroh180105Skill

    claude-handoff

    Hand the current conversation off to a fresh background agent that picks up the work immediately.

    Also in 41 projects that use this

  • LesaniSkill

    writing-shape

    Take a markdown file of raw material and shape it into an article through a conversational session — drafting candidate openings, growing the piece paragraph by paragraph, arguing about format (lists, tables, callouts, quotes) at each step. Use when the user has a pile of notes, fragments, or a rough draft and wants help turning it into something publishable.

    Also in 54 projects that use this

  • simo97Skill

    migrate-to-shoehorn

    Migrate test files from `as` type assertions to @total-typescript/shoehorn. Use when user mentions shoehorn, wants to replace `as` in tests, or needs partial test data.

    Also in 47 projects that use this

  • +1 more from Lesani

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 154 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Don't use it for what the agent can do: If the step is solved with a command, run it and move on.

    line 19
  2. Very specific:

    Done when: every stage has a name and an order, and for every value you know where the person gets it, where it is written (.env, secret, …

    line 52
  3. Distinctive:

    Done when: every stage is a set of instructions that someone who has never seen the project could follow.

    line 67
  4. Very specific:

    Use the helpers: stage, say/step, open_url, ask/ask_secret, write_env, set_secret/set_var, pause/confirm.

    line 75
  5. Distinctive:

    Keep the bar the template sets:

    line 78

“Common” is measured: kept by at least 6 in 10 of the 154 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

wizard

Some steps the agent cannot take: logging into a dashboard that requires a person's session, clicking "create" on something that costs money, reading a token that is only shown once. Those steps always end the same way: a list of instructions in the chat that is lost when the session closes, and two months later they have to be explained again.

A wizard turns that list into an executable script. It opens each URL, says exactly where to click and what to copy, collects the value, writes it where it goes (.env, GitHub secret), asks for confirmation before anything irreversible and shows how many stages remain.

Don't use it for what the agent can do. If the step is solved with a command, run it and move on.

What is already solved

template.sh ships the whole library: staged progress, confirmation gates, URL opening on Linux, macOS and WSL, hidden input for secrets, idempotent writing to .env, gh secret/gh variable and a final summary of what is still pending.

Your job is only to scope the procedure and write its stages. Everything above the STAGES marker is identical in every wizard, and that uniformity is the point: it is not edited by hand.

A wizard is disposable by default: it is written in the scratchpad or in scripts/, run, and deleted. It is committed only if the procedure will be repeated and deserves to live in the repo.

Procedure

1. Scope

Find out all the manual steps and all the values that come out along the way. Read the repo before asking, never start cold:

  • For configuration: .env, .env.example, README, docker-compose*, the framework config and .github/workflows/*. Every secrets.* and every vars.* in the workflow is a value the wizard has to produce.
  • For a migration or an infrastructure change: the current state, the final state and the irreversible actions between the two.

Show the user the ordered list of stages and which value each one produces, and have them confirm. They can add, remove or reorder.

Done when every stage has a name and an order, and for every value you know where the person gets it, where it is written (.env, secret, both or neither) and whether it is secret or public.

2. Trace the path of each stage

For each stage, write the exact route: which URL is opened, what is done there, where the value appears and which variable it fills. For example: "Dashboard → Developers → API keys → Reveal test key → copy".

If you don't know the current interface or the exact command, say so and ask, or look it up in the documentation. Inventing a menu that doesn't exist leaves the person stuck in front of a screen that doesn't look like what the script says, and then the wizard is worth less than nothing.

Done when every stage is a set of instructions that someone who has never seen the project could follow.

3. Write it

Copy template.sh to its destination. Replace the example stage with one stage per step, in dependency order, and adjust TOTAL_STAGES.

Use the helpers: stage, say/step, open_url, ask/ask_secret, write_env, set_secret/set_var, pause/confirm.

Keep the bar the template sets:

  • Open the URL before asking for the value found on it.
  • ask_secret for anything secret.
  • write_env for anything that has to persist.
  • set_secret only for what CI really needs.
  • confirm before any irreversible action.
  • One stage, one task. Each stage clears the screen, so if you put three things in one, the first is gone before the person reads it.

4. Check and hand over

bash -n <script>          # syntax
shellcheck <script>       # if installed
chmod +x <script>

Don't run it yourself from start to finish: it opens browsers and waits for a person. Review it by reading: every value from step 1 is collected and ends up where you said, and every set_secret name matches a secrets.* in the workflow exactly.

Tell the user how to launch it. If the procedure will be repeated, commit it and link it from the README, so the next person runs the script instead of asking an agent.


Adapted from the wizard skill of mattpocock/skills, MIT license. template.sh keeps the structure of the original library.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.