agentleFS
David-Florian2013David-Florian2013/Stop-trusting-made-it-faster-here-is-5-best-skills-for-Claude-/skills/prompt-clarifier/SKILL.md

prompt-clarifier

Skill1 starsChanged 41 days agoMITView on GitHub

Turns vague, underspecified coding-task requests into precise, actionable specifications before work begins. Use this skill whenever a user's request is short or ambiguous — missing a target file/function, missing success criteria, or using a vague verb like "fix this", "make it faster", "clean this up", "add a login feature", "improve this". Propose a clearer restatement of the same request and get it confirmed before starting. Do not trigger when the request already names what to change, where, and what success looks like.

2 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 1

    1. David-Florian2013/prompt-clarifier★ 091% same

    Wrote their own version · 1

    1. alunadev/ald-skills★ 4same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository5 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the request already names the target, the intended change, and how to verify it — skip this skill entirely and go straight to work.

    line 20
  2. Distinctive:

    Check the request against the four elements above.

    line 23
  3. Distinctive:

    Look for missing pieces in available context first (open files, error messages, repo structure, recent conversation) — many are implied by …

    line 24
  4. Very specific:

    Never fill an unknown piece with a placeholder like [fill in] inside a "clarified" restatement — a restatement with blanks in it isn't …

    line 27
  5. Distinctive:

    If the target is known (stated, or inferred from context) but other details are missing, draft a full clarified restatement of the *same* …

    line 28
  6. Distinctive:

    Wait for confirmation or an edit before doing any work.

    line 38

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “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

Prompt Clarifier

Why this matters

A vague task request forces guesswork: which file, what "faster" means, what counts as done. Guessing wrong wastes an entire work cycle — wrong file explored, wrong fix applied, more back-and-forth than if the request had been clear from the start. Clarifying up front costs one short exchange and saves everything downstream.

When to use this skill

Trigger when a request is missing two or more of:

  • Target — which file, function, module, or page is affected
  • Concrete change — what should actually happen, not just a verb ("fix", "improve", "clean up")
  • Success criteria — how to tell the work is done and correct
  • Scope/constraints — what should NOT change

When NOT to use it

If the request already names the target, the intended change, and how to verify it — skip this skill entirely and go straight to work. Re-clarifying an already-clear request is friction, not help.

Process

  1. Check the request against the four elements above. If it's already specific, do nothing and proceed normally.
  2. Look for missing pieces in available context first (open files, error messages, repo structure, recent conversation) — many are implied by what's in front of you even when not stated outright.
  3. If the target is still unknown after checking context, stop there: you can't honestly state success criteria or constraints for a file or page you can't identify. Ask directly and briefly for just the target, in plain language, while confirming what you do already understand about the intent:

    "Which page/file do you mean? I get that you want it faster — just need to know where." Never fill an unknown piece with a placeholder like [fill in] inside a "clarified" restatement — a restatement with blanks in it isn't clarified, it's the same question wearing a costume, and it reads as broken to whoever's reading it.

  4. If the target is known (stated, or inferred from context) but other details are missing, draft a full clarified restatement of the same request (never change what the user is asking for, only make it concrete):
    • Name the target
    • State the concrete change
    • State constraints / what must not change
    • State how success will be checked Show both versions and ask for confirmation:

    You said: "<original>" Here's a more specific version of the same request: <clarified> Proceed with this, or would you adjust anything?

  5. Wait for confirmation or an edit before doing any work. If the user corrects the clarified version, use their correction, not your own guess.

Keep it short

The clarified version should be 2-4 sentences, not a spec document. It restates intent precisely; it doesn't add new requirements the user didn't ask for.

Examples

Example 1 User: "oprav to" Clarified: "Oprav chybu v auth.js, kde přihlášení s platným heslem občas vrací 401 (vypadá to na race condition v ověření tokenu). Hotovo je, když existující login testy projdou a 10 opakovaných přihlášení po sobě neselže ani jednou."

Example 2 User: "udělej to rychlejší" Clarified: "Sniž dobu načtení /dashboard pod 1 s na desktopu (teď ~3 s) — zaměř se na N+1 dotazy v dashboardController a zbytečné re-rendery v DashboardView. Nezasahuj do API kontraktu. Hotovo je, když Lighthouse performance skóre vzroste aspoň o 20 bodů."

Example 3 User: "přidej login" Clarified: "Přidej přihlašovací formulář (email + heslo) na /login, napojený na authService.login(). Po úspěchu přesměruj na /dashboard, při chybě zobraz srozumitelnou hlášku. Hotovo je, když projde celý flow end-to-end a existuje test na neplatné přihlašovací údaje."

Example 4 — target unknown, ask for just that User: "tahle stránka je pomalá, mrkni na to" (No open file or prior context points to a specific page.) Response: "Kterou stránku myslíš — soubor nebo route? Zbytek je jasný: chceš ji zrychlit, bez změny chování."

Example 5 — English User: "fix this" Clarified: "Fix the 401 on POST /login in auth.js when a valid password is sent (looks like a token-refresh race). Done when existing login tests pass and 10 back-to-back logins succeed."

Example 6 — English User: "make it faster" Clarified: "Cut /dashboard load time under 1s on desktop (currently ~3s) by fixing N+1 queries in dashboardController and extra re-renders in DashboardView. Do not change the API contract. Done when Lighthouse performance rises by at least 20 points."

Example 7 — English User: "add login" Clarified: "Add an email+password form on /login wired to authService.login(). Redirect to /dashboard on success; show a clear error on failure. Done when the flow works end-to-end and a test covers invalid credentials."

Example 8 — English, target unknown User: "this page is slow, look at it" (No open file or prior context points to a specific page.) Response: "Which page — file or route? The rest is clear: you want it faster, without changing behaviour."

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.