telegram-context

What This Skill Does

The telegram-context skill provides essential conversational continuity for OpenClaw agents operating within Telegram. By automating the retrieval of historical message data, the agent can understand the context of ongoing discussions even if the user reboots the session or connects from a new environment. This bridges the gap between disconnected sessions by leveraging the built-in message tool to pull recent interaction logs directly from the active chat, ensuring the AI model is briefed on the latest topics before the user begins typing. It is a configuration-free, toggle-based utility that respects session isolation.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal or agent interface:

clawhub install openclaw/skills/skills/fourthdensity/telegram-context

Once installed, the skill will begin monitoring the memory/telegram-context.json file. Ensure your OpenClaw gateway is correctly authenticated with your Telegram account, as this skill relies on the underlying message tool permissions to perform its fetch operations.

Use Cases

• Long-term Project Management: Keep track of requirements discussed in a Telegram group over several days without repeating yourself.
• Support Continuity: Allow the agent to reference previous troubleshooting steps sent by the user in the same chat.
• Asynchronous Collaboration: Use the agent as a "catch-up" tool that summarizes the messages sent since the last time you checked your Telegram notifications.

Example Prompts

1. "/telegram-context on"
2. "/telegram-context fetch 50"
3. "/telegram-context status"

Tips & Limitations

• Security First: Because this skill injects history into the LLM context, be mindful that sensitive information included in chat logs will be transmitted to your configured AI provider. If the conversation moves to private, sensitive financial or personal data, use /telegram-context off immediately.
• Context Windows: Avoid setting high fetch counts for extremely active groups, as this might consume your token limit quickly. A standard fetch count of 20 is usually sufficient for most conversational threads.
• Scope: Note that this skill only accesses the current chat. It cannot aggregate data across different Telegram channels or private direct messages outside of the one it is currently active in. Always verify the status of the skill using /telegram-context status before assuming the agent is aware of your latest message history.