Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Automate Telegram tasks via Rube MCP (Composio): send messages, manage chats, share photos/documents, and handle bot commands. Always search tools first for current schemas.
.claude/skills/davepoon-telegram-automation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | 114% | 0% |
| case-21 | ✓→✓ | = Same ✓ | 127% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 96% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 141% | 0% |
| case-06 | ✓→✓ | = Same ✓ | 301% | 0% |
Automate Telegram operations through Composio's Telegram toolkit via Rube MCP.
Toolkit docs: composio.dev/toolkits/telegram
RUBE_MANAGE_CONNECTIONS with toolkit telegramRUBE_SEARCH_TOOLS first to get current tool schemasGet Rube MCP: Add https://rube.app/mcp as an MCP server in your client configuration. No API keys needed — just add the endpoint and it works.
RUBE_SEARCH_TOOLS respondsRUBE_MANAGE_CONNECTIONS with toolkit telegramWhen to use: User wants to send text messages to a Telegram chat
Tool sequence:
TELEGRAM_GET_ME - Verify bot identity and connection Prerequisite]TELEGRAM_GET_CHAT - Get chat details and verify access Optional]TELEGRAM_SEND_MESSAGE - Send a text message Required]Key parameters:
chat_id: Numeric chat ID or channel username (e.g., '@channelname')text: Message text contentparse_mode: 'HTML' or 'MarkdownV2' for formattingdisable_notification: Send silently without notification soundreply_to_message_id: Message ID to reply toPitfalls:
_*[]()~>#+-=|{}.!<b>, <i>, <code>, <pre>, <a>When to use: User wants to share images or files in a Telegram chat
Tool sequence:
TELEGRAM_SEND_PHOTO - Send an image Optional]TELEGRAM_SEND_DOCUMENT - Send a file/document Optional]Key parameters:
chat_id: Target chat IDphoto: Photo URL or file_id (for SEND_PHOTO)document: Document URL or file_id (for SEND_DOCUMENT)caption: Optional caption for the mediaPitfalls:
When to use: User wants to get chat information or manage chat settings
Tool sequence:
TELEGRAM_GET_CHAT - Get detailed chat information Required]TELEGRAM_GET_CHAT_ADMINISTRATORS - List chat admins Optional]TELEGRAM_GET_CHAT_MEMBERS_COUNT - Get member count Optional]TELEGRAM_EXPORT_CHAT_INVITE_LINK - Generate invite link Optional]Key parameters:
chat_id: Target chat ID or usernamePitfalls:
When to use: User wants to modify or remove previously sent messages
Tool sequence:
TELEGRAM_EDIT_MESSAGE - Edit a sent message Optional]TELEGRAM_DELETE_MESSAGE - Delete a message Optional]Key parameters:
chat_id: Chat where the message is locatedmessage_id: ID of the message to edit or deletetext: New text content (for edit)Pitfalls:
When to use: User wants to forward messages or retrieve recent updates
Tool sequence:
TELEGRAM_FORWARD_MESSAGE - Forward a message to another chat Optional]TELEGRAM_GET_UPDATES - Get recent bot updates/messages Optional]TELEGRAM_GET_CHAT_HISTORY - Get chat message history Optional]Key parameters:
from_chat_id: Source chat for forwardingchat_id: Destination chat for forwardingmessage_id: Message to forwardoffset: Update offset for GET_UPDATESlimit: Number of updates to retrievePitfalls:
When to use: User wants to set or update bot command menu
Tool sequence:
TELEGRAM_SET_MY_COMMANDS - Set the bot's command list Required]TELEGRAM_ANSWER_CALLBACK_QUERY - Respond to inline button presses Optional]Key parameters:
commands: Array of command objects with command and descriptioncallback_query_id: ID of the callback query to answerPitfalls:
From username:
1. Use '@username' format as chat_id (for public channels/groups)
2. For private chats, numeric chat_id is required
3. Call GET_CHAT with username to retrieve numeric IDFrom GET_UPDATES:
1. Call TELEGRAM_GET_UPDATES
2. Extract chat.id from message objects
3. Use numeric chat_id in subsequent callsparse_mode: 'HTML' for <b>bold</b>, <i>italic</i>, <code>code</code>parse_mode: 'MarkdownV2' for *bold*, _italic_, code _ * [ ] ( ) ~ > # + - = | { } . !Bot Permissions:
Rate Limits:
Chat Types:
Message Limits:
| Task | Tool Slug | Key Params | |------|-----------|------------| | Verify bot | TELEGRAM_GET_ME | (none) | | Send message | TELEGRAM_SEND_MESSAGE | chat_id, text, parse_mode | | Send photo | TELEGRAM_SEND_PHOTO | chat_id, photo, caption | | Send document | TELEGRAM_SEND_DOCUMENT | chat_id, document, caption | | Edit message | TELEGRAM_EDIT_MESSAGE | chat_id, message_id, text | | Delete message | TELEGRAM_DELETE_MESSAGE | chat_id, message_id | | Forward message | TELEGRAM_FORWARD_MESSAGE | chat_id, from_chat_id, message_id | | Get chat info | TELEGRAM_GET_CHAT | chat_id | | Get chat admins | TELEGRAM_GET_CHAT_ADMINISTRATORS | chat_id | | Get member count | TELEGRAM_GET_CHAT_MEMBERS_COUNT | chat_id | | Export invite link | TELEGRAM_EXPORT_CHAT_INVITE_LINK | chat_id | | Get updates | TELEGRAM_GET_UPDATES | offset, limit | | Get chat history | TELEGRAM_GET_CHAT_HISTORY | chat_id | | Set bot commands | TELEGRAM_SET_MY_COMMANDS | commands | | Answer callback | TELEGRAM_ANSWER_CALLBACK_QUERY | callback_query_id |
Powered by Composio
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 4,623 | 6,320 | +37% | 1 | 1 | 0% | 445 | 2,558 | +475% | 0 | 0 | — |
case-02 | fail→fail | 5,759 | 3,100 | -46% | 1 | 1 | 0% | 1,106 | 2,558 | +131% | 0 | 0 | — |
case-21 | pass→pass | 8,315 | 7,774 | -7% | 1 | 1 | 0% | 1,550 | 3,526 | +127% | 0 | 0 | — |
case-03 | fail→fail | 6,438 | 5,860 | -9% | 1 | 1 | 0% | 1,317 | 2,524 | +92% | 0 | 0 | — |
case-04 | pass→pass | 9,037 | 5,066 | -44% | 1 | 1 | 0% | 1,591 | 3,116 | +96% | 0 | 0 | — |
case-05 | pass→pass | 8,283 | 7,160 | -14% | 1 | 1 | 0% | 1,518 | 3,660 | +141% | 0 | 0 | — |
case-06 | pass→pass | 4,338 | 5,792 | +34% | 1 | 1 | 0% | 818 | 3,284 | +301% | 0 | 0 | — |
case-07 | pass→pass | 8,817 | 6,958 | -21% | 1 | 1 | 0% | 1,351 | 3,350 | +148% | 0 | 0 | — |
case-08 | pass→pass | 9,017 | 3,992 | -56% | 1 | 1 | 0% | 1,570 | 2,818 | +79% | 0 | 0 | — |
case-09 | fail→fail | 5,327 | 5,929 | +11% | 1 | 1 | 0% | 1,166 | 2,615 | +124% | 0 | 0 | — |
case-10 | pass→pass | 6,537 | 4,720 | -28% | 1 | 1 | 0% | 1,004 | 2,970 | +196% | 0 | 0 | — |
case-11 | pass→pass | 6,067 | 6,166 | +2% | 1 | 1 | 0% | 1,141 | 3,120 | +173% | 0 | 0 | — |
case-12 | fail→fail | 7,223 | 6,207 | -14% | 1 | 1 | 0% | 1,265 | 3,288 | +160% | 0 | 0 | — |
case-13 | fail→fail | 5,205 | 5,793 | +11% | 1 | 1 | 0% | 933 | 2,679 | +187% | 0 | 0 | — |
case-14 | fail→fail | 3,994 | 4,346 | +9% | 1 | 1 | 0% | 766 | 2,414 | +215% | 0 | 0 | — |
case-15 | pass→pass | 11,441 | 6,260 | -45% | 1 | 1 | 0% | 1,931 | 3,314 | +72% | 0 | 0 | — |
case-22 | pass→pass | 13,883 | 11,713 | -16% | 1 | 1 | 0% | 2,727 | 4,537 | +66% | 0 | 0 | — |
case-16 | pass→pass | 7,652 | 4,893 | -36% | 1 | 1 | 0% | 1,356 | 3,012 | +122% | 0 | 0 | — |
case-17 | pass→pass | 7,782 | 6,007 | -23% | 1 | 1 | 0% | 1,430 | 3,340 | +134% | 0 | 0 | — |
case-18 | fail→pass | 7,657 | 3,601 | -53% | 1 | 1 | 0% | 1,330 | 2,846 | +114% | 0 | 0 | — |
case-19 | pass→pass | 3,779 | 3,023 | -20% | 1 | 1 | 0% | 562 | 2,565 | +356% | 0 | 0 | — |
case-20 | pass→pass | 10,861 | 10,260 | -6% | 1 | 1 | 0% | 2,204 | 4,260 | +93% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +5 percentage points is the difference between those two pass rates over the 18 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.