AI Agent Hub
Back to skills
Feishu Group Chat Bot LLM Integration icon

Feishu Group Chat Bot LLM Integration

AI Agent Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please install @user_e9b2a228/pap according to https://skillhub.cn/install/skillhub.md.

About this skill

Problem it addresses

Connecting an LLM to a Feishu group chat can get stuck on app permissions, the target chat_id, message polling or callback handling, tool-call parameters, and stray reply artifacts. This skill packages those steps into a bot.js project, making it easier for engineers to put a model into a specific group.

How it works

After startup, the bot reads group messages and calls the configured MODEL_NAME, LLM_BASE_URL, and LLM_API_KEY. Built-in skills include save_memory, load_memory, web_search, get_weather, and wiki_search. New capabilities can be added under skills/ by exporting { name, description, parameters, execute }, then reloading the bot. BOT_IDENTITY controls reply style, and model changes require tool-call support.

Boundaries and notes

It depends on a Feishu enterprise self-built app with im:message read and im:message:send_as_bot send permissions, plus a correct CHAT_ID. If the model does not support function calling, tools may not trigger. Check msg.msg_type handling, token refresh, and the axios dependency via logs. Use bot.log for normal output and bot.err for errors.

Use Cases

  • Ask the bot in a Feishu group to answer real-time questions using `web_search`.
  • Store key facts with `save_memory`, then retrieve them later via `load_memory`.
  • Have the bot query city weather or Wikipedia entries and reply directly in chat.
  • Add a custom function under `skills/`, export `execute`, reload, and call it in the group.

Best For

  • Backend engineers integrating Feishu bots, who need to configure app credentials, `chat_id`, and LLM endpoints.
  • Product engineers adding a Q&A entry to team groups, using the model for weather, wiki, and web search.
  • AI engineers maintaining tool-call chains, who need to extend skills and confirm tool-call support.
  • Ops engineers debugging Feishu bot failures, checking `bot.log` and `bot.err` for 401, dependency, or field issues.