
telegram-bot-development-skill
Telegram bot development skill for Claude Code, Codex, OpenCode and Cursor: build, review and debug Telegram bots and Mini Apps against the live Bot API, che…
Install with your AI
Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.
Install and set up telegram-bot-development-skill (release-binary project) into my current project. Found on https://claudeers.com/telegram-bot-development-skill Repo: https://github.com/nyxandro/telegram-bot-development-skill Homepage/docs: — Detected install method: release-binary → inspect the README Category: plugins. Platforms: cli, api. Read the repo's README for exact setup and env vars, then install it and wire it into my project. Claudeers Health Verdict: unknown; community-verified: false. Confirm the source before running anything.
Grab the latest release asset from GitHub.
# download a build from https://github.com/nyxandro/telegram-bot-development-skill/releases
git clone https://github.com/nyxandro/telegram-bot-development-skill
// compatibility
| Platforms | cli, api |
|---|---|
| Operating systems | — |
| AI compatibility | claude |
| License | MIT |
| Pricing | open-source |
| Language | Python |
Telegram Bot Development
A Claude Code skill that checks the live Telegram Bot API and your installed library before your AI agent writes bot code.
English · Русский
Build, review and debug Telegram bots and Mini Apps with Claude Code, Codex, OpenCode, Cursor or any other agent that reads Agent Skills. The skill works with any language and framework, from grammY, Telegraf, aiogram, python-telegram-bot and GramIO to Go, Rust, PHP and .NET libraries or raw HTTPS. The official docs decide what exists. The library installed in your project decides what your code can call.
Why
The Bot API ships a release about once a month, and a release can undo what the previous one
added. In July 2026 Bot API 10.2 added receiver_user_id to sendMessage. Six weeks later 10.3
replaced it with ephemeral_message_parameters. An agent that learned the API before August will
confidently write code with a parameter that is already gone.
Writing from memory goes wrong in three ways.
- The agent calls parameters and methods that no longer exist.
- It builds workarounds for things the API now does natively.
- It quotes limits and fees from a year ago.
This skill makes the agent read the current docs and search your installed library before it writes, reviews or recommends any Bot API call.
Install
| Agent | Command |
|---|---|
| Claude Code, as a plugin | /plugin install telegram-bot-development --marketplace nyxandro/skills |
| Claude Code, Codex, OpenCode, Cursor and others | npx skills add nyxandro/telegram-bot-development-skill |
| Any agent, by hand | Download the latest .skill archive and unpack it into the agent's skills folder |
Pick one. If you install it two ways, Claude Code loads the skill twice.
More install options and requirements
- The one-line plugin install needs Claude Code 2.1.275 or later. On an older version, add the
catalog first with
/plugin marketplace add nyxandro/skillsand then run/plugin install telegram-bot-development@nyxandro. - The catalog pins no version, so every commit here reaches plugin users as an update. Run
claude plugin marketplace update nyxandroandclaude plugin update telegram-bot-development@nyxandro, or turn on auto-update under/plugin→ Marketplaces. - With
npx skills, add-gto install the skill for every project. To install a given version, add its tag, as innyxandro/[email protected]. - You can also clone this repository and link the
telegram-bot-development/folder into your agent's skills folder. There is no build step. - The scripts need Python 3.8 or newer and access to core.telegram.org and telegram.org. On
Windows the interpreter is usually
py -3orpython. On macOS with Python from python.org, runInstall Certificates.commandonce, or HTTPS requests fail. Without Python the agent reads the same official pages as plain URLs. - Tested on Linux and Windows.
Usage
Talk about your bot as usual. The skill starts whenever code, a question or a plan touches a Telegram bot, the Bot API, BotFather or a Mini App, even for a small change like a new button.
- My aiogram 3 bot answers through an LLM and takes about 20 seconds. Show the answer as it's generated and add a Stop button.
- In a group, show the rules only to the member who pressed the Rules button. Can our grammY version even do that?
- Review my webhook for reliability and security before launch.
- We sell a subscription for Telegram Stars. What do we turn on in BotFather, and what will it cost us?
How it works
flowchart LR
A[Your task] --> B[Latest Bot API and changes]
B --> C[Pages of every method and type]
C --> D[Your installed library]
D --> E[Code or review]
E --> F[What was checked]
- The agent finds the current Bot API version and reads every change since the version your library declares.
- It reads the official page of every method and type it touches, and the history of any name it isn't sure is still alive.
- It checks that your installed library accepts that parameter on that exact method. A name that merely appears somewhere in the library doesn't count, and neither does an old parameter the library still accepts as deprecated.
- Before it says "Telegram can't do that", it reads the features page and searches the docs.
- It ends by saying what it checked, the Bot API version and date, and whether your library supports what it used.
Example
This is real output of the skill's docs script, shortened. It's what the agent sees before it
writes a single sendMessage call.
$ python3 scripts/tgdocs.py history receiver_user_id
10.3 (August 24, 2026): Added the class EphemeralMessageParameters and replaced the parameters
receiver_user_id and callback_query_id in sendMessage, sendPhoto, sendVideo ... with the
parameter ephemeral_message_parameters.
10.2 (July 14, 2026): Added the parameters receiver_user_id and callback_query_id to sendMessage,
sendPhoto, sendVideo ...
Then it checks your project with find_method_support.py sendMessage.ephemeral_message_parameters
and writes code only with what both the docs and your library support.
What it covers
Eleven references load only when a task needs them.
| Topic | When the agent reads it |
|---|---|
| Staying current | Any method call, a "this doesn't exist" report, a library that lags behind the API |
| Receiving updates | Polling or webhooks, deploys, scaling, shutdown, duplicate or lost updates |
| Conversations | /start, menus, keyboards, callbacks, commands, formatting, long text, deep links, groups and topics |
| Security and privacy | The token, the webhook secret, sign-in, permissions, user data, the privacy policy, abuse |
| Limits and performance | Broadcasts, bursts, 429 errors, files, throughput |
| Errors and edge cases | Error handling, retries, flaky behaviour, odd updates |
| Product features | What a bot can do today, product options, BotFather settings |
| Stars payments | Selling, subscriptions, refunds, payment disputes |
| Mini Apps | Launch, sign-in, theme, buttons, storage |
| AI bots | Streaming answers, formatting model output, threads, bot-to-bot, guest mode |
| Review checklist | A review or audit before release |
The developer terms are read clause by clause, including the fee on Stars purchases in bots that enable topics in private chats.
Tools
Two scripts, Python 3.8 or newer, standard library only. Pages are cached for six hours in
~/.cache/telegram-bot-docs.
| Script | What it does |
|---|---|
scripts/tgdocs.py | The official docs in their Markdown form. The latest version, changes since a version, a method or type page, search, the history of a name, any guide and the developer terms |
scripts/find_method_support.py | Finds the Telegram libraries installed in a project and shows whether each one knows a method, a type or a METHOD.PARAMETER, with the matching lines |
Versions
The current version sits in metadata.version of telegram-bot-development/SKILL.md, so any
installed copy tells you which one it is. Every version has a git tag and a
release with its own
archive. CHANGELOG.md lists what changed.
Tests
uv run --with pytest --with hypothesis python -m pytest test
The tests run offline. Small fake pages and packages stand in for the docs and the libraries. pytest and Hypothesis are needed only for the tests, not by the skill.
telegram-bot-development/evals/ holds test tasks for the agent, their assertions and the results
of past runs. Its README describes the fixtures. The folder is left out of the archive.
Repository layout
telegram-bot-development/ the skill
SKILL.md the workflow, live API and installed library first
references/ eleven references, read only when a task needs them
scripts/tgdocs.py the official docs, live
scripts/find_method_support.py what the project's installed library knows
evals/ test tasks for the agent and their results, not in the archive
test/ offline tests of both scripts
dist/telegram-bot-development.skill the skill folder as a downloadable archive
CHANGELOG.md what changed in each version
README.md, README.ru.md this page in English and in Russian
License
// faq
What is telegram-bot-development-skill?
Telegram bot development skill for Claude Code, Codex, OpenCode and Cursor: build, review and debug Telegram bots and Mini Apps against the live Bot API, checking the official docs and your installed library (aiogram, grammY, python-telegram-bot, Telegraf) before any call.. It is open-source on GitHub.
Is telegram-bot-development-skill free to use?
telegram-bot-development-skill is open-source under the MIT license, so it is free to use.
What category does telegram-bot-development-skill belong to?
telegram-bot-development-skill is listed under plugins in the Claudeers registry of Claude-compatible tools.
// embed badge
[](https://claudeers.com/telegram-bot-development-skill)
// retro hit counter
[](https://claudeers.com/telegram-bot-development-skill)
// reviews
// guestbook
// related in Claude Plugins
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explainin…
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/
financial-services — a Claude ecosystem project on GitHub.