Store Submit logo

Store Submit

CommunityPopular
zhitongblog
store-submit

Publish a SoloMD release to the stores that have no usable submission API — Google Play Console and Microsoft Partner Center — by driving them through the local Unzoo Browser REST API. Covers upload, release notes, track and country selection, and submitting for review, plus the traps that have cost whole sessions before. Use when asked to "submit to Play", "上架 Play", "update the Microsoft Store", "提交商店", "发到应用商店", or when a release has shipped on GitHub and the store channels are still owed. Apple is NOT here: the App Store has a real API — see scripts/submit-for-review.sh.

Overview

Publisherzhitongblog
Repositorysolomd
Skill namestore-submit
Stars
1.2K
Forks
80
Bundled files
3
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by zhitongblog on GitHub. Read the source before you install it.

Installation

Install the Store Submit AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/zhitongblog/solomd/tree/main/.claude/skills/store-submit
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/zhitongblog/solomd.git /tmp/solomd
mkdir -p .claude/skills
cp -r /tmp/solomd/.claude/skills/store-submit .claude/skills/store-submit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Store Submit in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Store Submit on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Store Submit is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Store submission over the Unzoo browser

Google Play and Microsoft Partner Center are the two channels with no usable API for this account, so releasing to them means driving a real browser. This skill is the accumulated way to do that without losing an afternoon.

Apple does not belong here. iOS and macOS submit over the App Store Connect API with no browser at all:

bash
./scripts/submit-for-review.sh --platform ios   --version X.Y.Z \
    --notes-dir app-store/release-notes/X.Y.Z --dry-run

Reach for the browser on Apple only if that script has failed for a reason you have actually diagnosed.

Preflight

bash
S=.claude/skills/store-submit
$S/unzoo.sh status          # {"browser_connected":true,...}
$S/unzoo.sh up              # starts the browser if it is down
$S/unzoo.sh tabs            # find the console tab that is already signed in

If the daemon is down, start it from a Bash call with run_in_background: true — nohup … & is not enough, it dies with the tool's shell:

"/Applications/Unzoo Browser.app/Contents/MacOS/unzoo-service" --daemon

browser_connected: false while the daemon is up means every navigate and evaluate will return HTTP 500; unzoo.sh up opens the browser and waits.

The six rules

These are not style preferences. Each one has already cost a session.

1. Pin the tab. Always. Without tab_id the API acts on whatever tab the user is looking at — their mail, their bank. Resolve one and export it:

bash
export UNZOO_TAB=$($S/unzoo.sh find-tab play.google.com/console)

If a read comes back nonsensical, check unzoo.sh url before you debug the page. It has been the user's Cloudflare dashboard.

2. A coordinate click goes to the ACTIVE tab, not to tab_id. The endpoint accepts the id and ignores it. Clicking a background tab produces no click event, no error, and a page that looks like it ignored you. unzoo.sh click and click-el activate first for this reason — if you call /api/v1/click directly, activate yourself.

3. Read pages with text, never with screenshots. A screenshot does not follow scroll, so full_page silently omits exactly the banner you are looking for. Use a screenshot only to locate something you must click by coordinate.

4. Use the console's own session, in the user's own profile. The signed-in Google and Microsoft sessions live in the default profile. Prefer find-tab over new. Do not touch cookies: setting one without SameSite=None has signed the main profile out.

5. Never claim a submission you have not read back. Both consoles render optimistically and will lie in both directions. Partner Center showed "Update in certification" seconds after the click, then "Update in draft" in a screenshot taken moments later, with a session-expired dialog in the DOM throughout — and the submission had in fact gone through. Navigate away and back before reading a status, and look for a structural change (the Submit button gone, a Cancel button appeared), not just a word.

6. A file input that reads back empty was probably filled. Angular and React uploaders take the FileList and clear the input in the same tick, so a successful set_input_files leaves files.length === 0. Confirm on the page. Retrying on the count uploads twice.

Which store

  • Google Play → play.md. AAB upload, release notes, countries, review.
  • Microsoft Store → msstore.md. MSIX packaging, Partner Center, certification.

unzoo.sh

status | up | tabs | find-tab <url-substring> | new <url> | activate
nav <url> | url | wait-load | wait-for <sel> [s] | exists <sel>
text | html | eval <js> | box <sel> | click <x> <y> | click-el <sel>
upload <input-sel> <file> | shot [path]

-t <id> overrides UNZOO_TAB for one call. Full API surface, if you need something not wrapped here:

bash
curl -s --noproxy '*' http://127.0.0.1:9399/api/v1/openapi.json

Note the --noproxy '*': the local proxy mangles loopback calls, and it is also why store uploads through the browser stall — see reference_release_upload_watchdog in memory.

★ text lies when the window is not on screen

unzoo.sh text and any innerText read need layout, and Chrome skips layout for a browser window that is not actually being painted. The Play Console page then reads as ~550 characters of loading splash while it is fully rendered — a screenshot shows the real content. This is silent: no error, just a short string that looks like "still loading".

Anything unattended (a watcher, a poll loop, a cron) must read unzoo.sh dom-text, which returns document.body.textContent and does not depend on layout. Two caveats that come with it:

  • textContent includes hidden and collapsed elements, so match on markers you expect to appear; do not conclude something is gone because a string is absent from a freshly loaded page.
  • Loading-splash text stays in textContent permanently, so judge "page ready" by length plus a content anchor, never by the absence of the splash string.

When a read disagrees with what you think is on screen, take a screenshot — that forces a paint and settles it.

★ eval used to share one scope, and a collision looked like null

Every unzoo.sh eval landed in the same top-level context, so the second const b = ... in a session threw "Identifier 'b' has already been declared" — and that throw came back as a plain null result, indistinguishable from "the selector matched nothing". It sent this session chasing imaginary DOM changes more than once, and it is visible in the page console as a stray SyntaxError that has nothing to do with the page.

eval now wraps every expression in its own arrow scope, so redeclaration is impossible. Three input shapes are handled: a bare expression, several statements where you write your own return, and several statements where you do not (the last one is returned for you). A ; inside a string literal is the one case that misfires — and it fails as a visible syntax error, not silence.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Store Submit AI skill do?

Publish a SoloMD release to the stores that have no usable submission API — Google Play Console and Microsoft Partner Center — by driving them through the local Unzoo Browser REST API. Covers upload, release notes, track and country selection, and submitting for review, plus the traps that have cost whole sessions before. Use when asked to "submit to Play", "上架 Play", "update the Microsoft Store", "提交商店", "发到应用商店", or when a release has shipped on GitHub and the store channels are still owed. Apple is NOT here: the App Store has a real API — see scripts/submit-for-review.sh.

Why use Store Submit on TypingMind?

Because you install it once and use it with any model. Store Submit is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Store Submit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/zhitongblog/solomd/tree/main/.claude/skills/store-submit. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Store Submit?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Store Submit?

As many as you like. As long as a model supports skills, you can use Store Submit with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Store Submit AI skill free?

Yes. It is published on GitHub by zhitongblog under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇