telegram-launch
Skill0 starsChanged 36 days agoMITView on GitHub
Get a Telegram bot or Mini App ready for real users — release checklist for secrets, webhooks, payments, Mini App security and TON parts, hosting and monitoring, BotFather listing and profile, growth and monetization options and their rules. Use this skill whenever the user is about to launch, release or publish a bot or Mini App, asks "is it ready", "what am I missing", how to promote or monetize a bot, or wants a production checklist.
- Reads credentials
In this repository11 other files it gives its agents.
- add-paymentsSkill
- add-ton-connectSkill
- audit-projectSkill
- deploy-testnetSkill
- new-botSkill
- new-mini-appSkill
- telegram-botSkill
- telegram-devSkill
- telegram-mini-appSkill
- telegram-paymentsSkill
- telegram-tonSkill
The file
Files
Launching in Telegram
Go through the sections that apply, report each item as done, missing or not applicable, and
ask before changing code. For an automated pass run /telegram-dev:audit-project.
Secrets and access
- Bot token, API keys and webhook secret only in the host's env vars;
.envgit-ignored;.env.examplewith placeholders. - Nothing secret in the Mini App bundle (search the built files, not only the source).
- The Telegram account that owns the bot has 2-step verification.
- A plan for a leaked token: revoke in @BotFather, update env, redeploy.
Bot
- Webhook with
secret_token, checked on every request;allowed_updateslists every update type the code handles. - Handlers idempotent and quick to answer; slow work in a queue.
- 429 handled with
retry_after; broadcasts throttled and resumable; blocked users (403) marked and skipped. /startcopes with missing, unknown and old deep-link parameters.- Tested in a private chat, a group (privacy mode), and the Telegram test environment.
- Commands set with
setMyCommands, description and about text, profile photo.
Mini App
initDatavalidated on the server, with anauth_datelimit.- Works under the cross-origin restriction (no Mini App calls from embedded foreign pages).
- Theme, safe areas and full-screen checked on iOS, Android, Desktop and Web.
- Loading and error states; works on slow mobile networks.
Payments
- Digital goods in Stars only. No crypto or card checkout for them.
- Pre-checkout answered within 10 seconds; fulfilment only on
successful_payment, unique ontelegram_payment_charge_id. /termsand/paysupportwork; refunds tested withrefundStarPayment.- Payment records backed up; reconciliation with
getStarTransactions.
TON parts (if any)
- Network shown in the UI; testnet and mainnet configs can't be mixed up.
- Payments credited only from
finalized(or a documentedconfirmedchoice for cheap goods), with a checkpoint and catch-up after restarts. - Jetton masters allowlisted;
ton_proofverified on the server with a one-time nonce. - Contracts reviewed by the
ton-security-revieweragent; a human audit for anything that holds real value. Mainnet deployment is done by the user.
Hosting and operations
- HTTPS with a valid certificate; a health check; restarts on crash.
- Logs without personal data or secrets; alerts on error spikes and webhook failures
(
getWebhookInfoshowslast_error_messageand pending updates). - Database backups; a way to roll back a deploy.
- A privacy policy if you store any user data.
Growth and money
Options Telegram lists today (terms change, link the official page instead of promising numbers): Stars for digital goods and services, subscription plans, paid media, affiliate programs with custom revenue share, 50% of the revenue from Telegram Ads shown in the chat with the bot, and using earned Stars for higher message limits, gifts to users or rewards in Toncoin. Sources: https://core.telegram.org/bots/features#monetization and https://telegram.org/tos/bot-developers.
Promotion that doesn't get the bot banned: users opt in before you message them, every broadcast has a clear reason and an easy way to stop, no bought audiences, no userbot spam.
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
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.

