agentleFS
torisKRtorisKR/toris-agent/skills/app-store-listing-creator/SKILL.md

app-store-listing-creator

Skill2 starsChanged 4 months agoApache-2.0View on GitHub

Create or improve Google Play and Apple App Store listing packages, including app icon direction, screenshot storyboards and prompts, promotional copy, short and full descriptions, keyword strategy, localization notes, and submission preflight. Use when launching or relaunching an app, preparing store assets, improving listing conversion, localizing a listing, or auditing existing store creative and copy.

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. About 0 are common, 15 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈15

    Distinctive · estimate

    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
  • could be automatic: a setting could enforce it
  1. Very specific:

    Read execution defaults and select execute, review, or plan.

    line 12
  2. Distinctive:

    create an evidence inventory and store/locale matrix, verifying current official constraints when exact limits matter.

    line 15
  3. Distinctive:

    Follow the user's language for deliverables unless a target locale requires another language.

    line 22
  4. Distinctive:

    Verify current dimensions, file formats, character limits, field availability, and review policies against official Apple or Google …

    line 24
  5. Distinctive:

    Never invent awards, rankings, adoption numbers, testimonials, certifications, privacy claims, or capabilities.

    line 25
  6. Distinctive:

    Do not promise store approval or conversion improvement.

    line 26
  7. Distinctive:

    Read evidence and verification before evaluating claims or reporting expected outcomes.

    line 27
  8. Distinctive:

    Ask only for missing information that would materially change the result.

    line 40
  9. Very specific:

    Use current official store documentation.

    line 50
  10. Distinctive:

    Do not reuse an old remembered limit when official documentation is accessible.

    line 52
  11. Distinctive:

    Avoid text, screenshots, tiny detail, misleading badges, and direct imitation of a competitor.

    line 70
  12. Distinctive:

    If the user authorizes image creation and an image tool exists, generate the selected concept only after the direction is clear.

    line 72
  13. Distinctive:

    Treat screenshots as a sequence, not isolated posters:

    line 76
  14. Distinctive:

    Create only fields supported by the verified store matrix.

    line 87
  15. Distinctive:

    Read deliverables and produce its complete package.

    line 97
  16. Distinctive:

    When experimentation is possible, propose one-variable hypotheses with success and guardrail metrics.

    line 99
  17. Distinctive:

    Every final claim is supported, softened, or explicitly marked as missing evidence.

    line 108

≈ “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

App Store Listing Creator

Build an evidence-based listing package for Google Play, Apple App Store, or both. Produce actual images when an image-generation capability is available and authorized; otherwise produce generation-ready creative briefs and prompts.

Quick start

Read execution defaults and select execute, review, or plan. In the default execute mode:

  1. inspect the product, existing listing/assets, target stores/locales, and supplied evidence;
  2. create an evidence inventory and store/locale matrix, verifying current official constraints when exact limits matter;
  3. produce the complete positioning, icon direction, screenshot storyboard, copy, localization, and preflight package;
  4. generate requested visual concepts when an image tool is available and authorized;
  5. validate the package, but do not upload or publish it without explicit authorization.

Operating rules

  • Follow the user's language for deliverables unless a target locale requires another language.
  • Inspect supplied product evidence and existing assets before proposing copy or visuals.
  • Verify current dimensions, file formats, character limits, field availability, and review policies against official Apple or Google documentation before giving exact submission constraints.
  • Never invent awards, rankings, adoption numbers, testimonials, certifications, privacy claims, or capabilities.
  • Do not promise store approval or conversion improvement.
  • Read evidence and verification before evaluating claims or reporting expected outcomes.

Workflow

1. Establish the listing matrix

Identify:

  • stores, app type, package/bundle identifiers, release stage, and target date;
  • target countries, languages, devices, orientations, and form factors;
  • primary audience, job to be done, differentiators, pricing, and required disclosures;
  • available build, screenshots, icon/brand files, design tokens, analytics, reviews, and approved product facts.

Ask only for missing information that would materially change the result. If safe, continue with clearly labeled assumptions.

2. Audit existing material

For an existing listing, record each field and asset before editing. Evaluate message hierarchy, consistency with the product, first-impression clarity, legibility at store scale, localization quality, unsupported claims, duplicated frames, and gaps between screenshots and the live experience.

Separate observed evidence from hypotheses. Do not infer conversion causes from a single aggregate metric.

3. Verify current platform constraints

Use current official store documentation. Record the source and access date for volatile rules. Build a platform/locale/asset matrix containing required and optional fields, exact limits, formats, device sets, and policy-sensitive content.

Do not reuse an old remembered limit when official documentation is accessible.

4. Define positioning

Write:

  1. one primary audience statement;
  2. one core promise supported by product evidence;
  3. three to five proof-backed value pillars;
  4. a message hierarchy for the first impression, evaluation, and trust stages;
  5. keyword themes mapped to real intent rather than repeated mechanically.

Resolve conflicting messages before producing assets.

5. Create the icon direction

Produce two or three meaningfully different concepts. For each include the central metaphor, silhouette, composition, palette, depth/material treatment, small-size behavior, platform adaptation, brand linkage, risks, and a generation-ready prompt.

Avoid text, screenshots, tiny detail, misleading badges, and direct imitation of a competitor. Test recognition at small size, monochrome/background variants when relevant, and contrast against common store surfaces.

If the user authorizes image creation and an image tool exists, generate the selected concept only after the direction is clear. Preserve editable source or prompt details when possible.

6. Build the screenshot narrative

Treat screenshots as a sequence, not isolated posters:

  • Frame 1: instantly identify the product outcome.
  • Frames 2–3: show the strongest differentiators with real UI evidence.
  • Middle frames: explain workflow, depth, or trust.
  • Final frame: reinforce the next action or ecosystem value without unsupported urgency.

For every frame specify objective, headline, optional support line, source screen/state, crop, device treatment, layout, background, visual emphasis, localization risk, accessibility/legibility note, and generation or compositing prompt. Never fabricate UI that implies unavailable functionality.

7. Write store copy

Create only fields supported by the verified store matrix. Keep the opening specific, readable, and outcome-led. Explain features through user value, then include necessary trust, subscription, compatibility, or legal details.

Map keywords naturally to relevant fields. Do not stuff, repeat irrelevant terms, use competitor trademarks deceptively, or claim superlatives without evidence. Preserve mandatory legal and privacy language exactly when supplied.

8. Localize deliberately

For each locale, identify literal-translation risks, cultural or legal sensitivities, text expansion, typography, screenshots requiring replacement, keyword intent differences, and claims needing local evidence. Prefer transcreation guided by the positioning over word-for-word translation.

9. Assemble and verify

Read deliverables and produce its complete package. Check every fact against evidence, every asset against the current platform matrix, copy against verified limits, screenshots against the real product, and locale variants for truncation and meaning.

When experimentation is possible, propose one-variable hypotheses with success and guardrail metrics. Label them as experiments, not predicted wins.

Mutation boundary

Creating briefs, copy, and an audit is read-only. Editing design files, generating images, changing store-console content, or uploading assets requires explicit user authorization and the relevant access. Never submit or publish a listing merely because the user requested an audit or creation package.

Definition of done

  • The requested stores and locales have a complete, current field/asset matrix.
  • Every final claim is supported, softened, or explicitly marked as missing evidence.
  • Icon direction, screenshot sequence, store copy, localization notes, and preflight are production-ready.
  • Generated or proposed visuals match the real product and remain legible at target sizes.
  • Current platform constraints and all requested character/asset checks are verified.
  • The handoff follows the shared format and clearly identifies any upload or owner action still required.

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.