A specialized sub-agent skill for formatting and delivering rich, scannable Telegram messages with HTML styling and structured layouts.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install telegram-compose
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install telegram-compose using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Telegram Compose?
Telegram Compose is a sophisticated utility within the Openclaw Skills ecosystem designed to transform raw AI-generated text into professional-grade Telegram communications. While simple replies are handled by standard tools, any substantive content—such as reports, research summaries, or structured data—is routed through this skill to ensure high visual hierarchy and readability on mobile devices.
By leveraging the direct Telegram Bot API and specific HTML formatting rules, the skill provides a robust framework for delivering complex information. It handles everything from character escaping and mobile-friendly data display to the intelligent splitting of long messages, ensuring that your agent's output is always clear, scannable, and visually consistent.
Telegram Compose Use Cases
- Delivering structured research summaries and briefings with clear visual headers.
- Sending system status updates or alerts with emoji anchors and bold labels.
- Formatting complex data lists and cards that flow naturally on mobile screens.
- Automating the delivery of long reports by splitting them into sequential, valid HTML chunks.
How Telegram Compose Works
- The main session agent detects if a response is substantive (more than 3 lines or contains structured data).
- A sub-agent is spawned specifically for the Telegram Compose task using the Claude Haiku model.
- The skill extracts the bot account name, chat ID, and raw content from the task parameters.
- Formatting rules are applied, including the insertion of HTML tags for bolding, quotes, and expandable blocks.
- Special characters are escaped (targeting ampersands and angle brackets) to prevent API parsing failures.
- The skill retrieves the bot token from the local Openclaw Skills configuration file.
- Messages are dispatched via curl; if the content exceeds 4,096 characters, it is automatically split at section boundaries.
- The sub-agent monitors the API response and implements error recovery, such as retrying without HTML if formatting errors occur.
Telegram Compose Setup
To integrate this skill into your environment, ensure you have the required system dependencies and configuration set up.
Install the necessary binaries:
sudo apt-get install jq curl
Configure your bot accounts in ~/.openclaw/openclaw.json or ~/.openclaw/clawdbot.json:
{
"channels": {
"telegram": {
"accounts": {
"main": {
"botToken": "YOUR_BOT_TOKEN_HERE"
}
}
}
}
}
Telegram Compose Data Schema & Taxonomy
The skill organizes content using a specific visual hierarchy to maximize readability within Openclaw Skills workflows:
| Component | Markdown/HTML Pattern | Visual Purpose |
|---|---|---|
| Header | EMOJI <b>CAPS TITLE</b> |
Primary section anchor |
| Labels | <b>Label:</b> Value |
Key-value data pairing |
| Quotes | <blockquote>Text</blockquote> |
Summaries or highlights |
| Details | <blockquote expandable> |
Collapsible long-form content |
| Links | <a href="...">Text →</a> |
Actionable call-to-action |
| Lists | • Bullet point |
Unordered information |
Telegram Compose Advanced Features
- Intelligent Message Splitting: Automatically breaks messages exceeding 4,096 characters into sequential chunks while maintaining valid HTML structure.
- Mobile-Optimized Data Display: Avoids monospace code blocks in favor of flexible, emoji-based rows that prevent layout breaking on small screens.
- Automated Error Fallback: Detects HTML parsing errors and automatically falls back to plain-text delivery to ensure message arrival.
- Multi-Account Support: Dynamically selects bot tokens based on the account name provided by the calling agent.
- Thread-Aware Messaging: Supports sending content to specific forum topics or message threads via the thread ID parameter.
SKILL.md
Loading
Related Openclaw Skills
image-sprout
A project-based image generation tool for maintaining visual consistency across multiple AI-generated outputs.
Clawcierge
A comprehensive concierge platform providing AI agents with essential infrastructure for messaging, cloud storage, and human-in-the-loop support.
ClawSend Messaging Skill
A secure messaging protocol for AI agents featuring cryptographic signing, end-to-end encryption, and structured intent-based communication.
clawlychat
An API-driven social networking layer designed for AI agents to communicate and interact on a global timeline.
Invoice Generator
A professional PDF generation tool that converts structured JSON data into formatted invoices with automatic versioning.
Vikunja Fast
A streamlined interface to manage Vikunja tasks and projects using Openclaw Skills for automated workflows.
METADATA
Featured*
