Skip to content

kharmanskyi/open-steps / skills/os-ask-simple

os-ask-simple

Verified

ALWAYS invoke this skill before asking the user any technical question or offering options, and whenever they ask to be asked in plain words - "ask simple", "ask me simply", "ask me in plain words" - in any language. ALWAYS invoke it too when the user asks for your view on a technical choice: whether it is worth doing, more than the problem needs, or replaceable by something simpler, and which option you would take. Invoke it even when the code makes the answer look obvious: the checks are what make the answer more than a guess. Rewrites the question in plain words and always ends with one marked recommendation. A structural choice first passes six checks, shown as a table: effort now, simpler substitute, extra work later, lock-in, over-engineering, easy to undo. Doing nothing is always weighed.

FileKindSize
SKILL.mdskill md5.3 KB
references/01-sign-in-now-or-later.mdmarkdown2.8 KB

references/01-sign-in-now-or-later.md

View on GitHub
# Example 1 - three questions in one, none of them answerable

Source: a real session in my own analytics dashboard project, condensed. No
client or third-party material.

## What the agent asked

> Should I integrate Clerk with Next.js middleware for route-level protection,
> or defer auth and keep the deployment protection for now? If Clerk, do you
> want the hosted sign-in page or embedded components, and should I model
> organizations from the start in case you go multi-tenant later?

Three questions stacked in one message, each assuming the answer to the one
before it, all in the vendor's vocabulary.

## `os-ask-simple`

> Should the dashboard get real sign-in now, or stay behind the single
> password it has today?
>
> Why it matters: today one password guards everything; real sign-in gives
> each person their own access, which you can take away one at a time.
> What changes later: nothing gets harder by waiting - the sign-in work is the
> same size on the day a second person needs it.
> Easy to undo: yes - the gate is one file and one screen; take them out and
> the password is back.
>
> | Check | Answer |
> |---|---|
> | How long now | Roughly half a day to a day; the gate in front of every page, plus the sign-in screen in the app shell |
> | Simpler substitute | The password already there - it does this one job today |
> | Extra work for you later | Inviting and removing people by hand in the provider's dashboard |
> | Harder to change later | Accounts lock to the provider's user ids; switching providers means re-mapping them |
> | Over-engineering | Yes, today - one person uses this dashboard |
> | Easy to undo | Yes: the gate is one file and one screen; remove them and the password is back |
>
> My recommendation: change nothing until a second person needs access. The
> password does the one job sign-in would do, and the half-day is better spent
> on the dashboard itself.
>
> **Sign-in?**
> - Keep the password (Recommended) - nothing to build; add sign-in the day a second person needs it
> - Add real sign-in now - roughly half a day to a day, then inviting people by hand

## What the rewrite did

| Move | Where |
|---|---|
| Three questions became one | hosted-vs-embedded and multi-tenant only exist after a yes - asking them first buys careless answers (rule 5) |
| Vendor words became what the reader would see | "Clerk with middleware" → "real sign-in"; "deployment protection" → "the single password" |
| Doing nothing was weighed, and won | rule 1 - and the one line says why it won |
| The screen said over-engineering, plainly | one person today; the row exists to be answered yes sometimes |
| Recommended against the more interesting build | the integration is the fun option; that was a reason for suspicion, not preference (rule 6) |
| The picker carried the trade-offs | two options, one line each, the recommended one first and marked |
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall