A standardized library of skills, practical workflows, and integration guides for the Telegram MCP Server.
This documentation is designed for developers, agent integrators, and AI agents (such as Claude Code, Cursor, Codex, and Antigravity) seeking reliable, copy-paste-ready patterns to interact with Telegram via the Model Context Protocol.
In this repository, a Skill represents a structured, reproducible agent workflow combining one or more Telegram MCP tools to accomplish an end-to-end task safely. Each skill provides:
- Concrete tool calls with required parameters.
- Explicit safety and privacy guardrails (e.g., read-only execution, drafting without sending).
- Step-by-step prompt-to-response sequences.
- Practical failure modes and recovery procedures.
- How to Use Telegram MCP: Complete setup, authentication, environment configuration, MCP client connections (Claude Desktop, Cursor, generic stdio), server capabilities, and privacy architecture.
| Workflow | Focus / Primary Tools | Safety Level | Document |
|---|---|---|---|
| Summarize Unread Messages | list_chats(unread_only=True), get_messages |
Read-Only | summarize-latest-unread.md |
| Draft Replies Without Sending | get_message_context, save_draft (or text draft) |
Safe / Non-transmitting | draft-replies-without-sending.md |
| Triage & Action Items | list_chats, get_messages, set_contact_alias |
Read-Only | triage-and-action-items.md |
| Search Chat & Context Summarization | list_messages(search_query=...), get_message_context |
Read-Only | search-chat-and-summarize-context.md |
- Goal: Get an executive briefing of all unread messages across direct chats and channels without marking them as read.
- Guide: Follow
skills/examples/summarize-latest-unread.md.
- Goal: Review recent conversation context and produce a draft for human approval or place it in the Telegram client's input box as a draft.
- Guide: Follow
skills/examples/draft-replies-without-sending.md.
- Goal: Scan active conversations for urgent questions, blockers, and assigned tasks, categorized by urgency.
- Guide: Follow
skills/examples/triage-and-action-items.md.
- Goal: Search message history across or within specific chats and reconstruct surrounding conversational context.
- Guide: Follow
skills/examples/search-chat-and-summarize-context.md.
All skills in this directory adhere to a 10-section standard template:
- Skill Name
- Purpose
- When to Use / Use Cases
- Prerequisites / Configuration
- MCP Tools Used (names, parameters, notes)
- Inputs (agent/user inputs)
- Outputs (structured results)
-
Step-by-step Example (prompt
$\rightarrow$ tool calls$\rightarrow$ response) - Safety / Privacy Notes (allowlist, read-only mode, untrusted user content)
- Failure Modes & Recovery (auth errors, rate limits, access denied)