Token导航 LogoToken导航TokenDH.com
研究检索执行命令clawhub未标认证来源可访问clear审计通过

feishu-new-chat飞书新聊天

Agent Skill

feishu-new-chat 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

9,126

周安装

388

GitHub Stars

公开资料未说明

下载量

3,197
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:feishu-new-chat(飞书新聊天)
来源仓库:https://github.com/dxxbb/feishu-new-chat
安装命令:
openclaw skills install feishu-new-chat
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install feishu-new-chat

简介

在飞书主题组中创建新聊天并发送初始回复,支持@提及。

  • 适用于需要快速开启群聊讨论或引导对话开场的场景。
  • 通过 OpenClaw 调用飞书 API 创建主题并发送消息。
  • 安装前需确认机器人权限及是否涉及敏感操作。feishu-new-chat 属于研究检索类 Skill,可作为该场景下的辅助能力补充。
  • 注意检查维护状态与 API 变更对功能的影响。

SKILL.md

name
feishu-topic-spawn
description
Create a new topic in a Feishu topic-group, and optionally add the first in-thread reply (with optional @mention) by sending as the user. Use when the user says things like ‘开话题’, ‘/topic’, ‘开个新话题’, ‘在话题里回一条’, or asks to simulate a ChatGPT-style ‘new chat’ workflow inside a Feishu topic-group. In a Feishu topic-group, a top-level send creates the topic; replying with reply_in_thread=true posts inside that topic.

Feishu Topic Spawn

Create a new Feishu topic in a dedicated topic-group, then optionally post the first thread reply inside it.

Keep this skill lightweight. Use it for topic creation and first-reply bootstrapping, not for session routing magic.

Validated behavior

This workflow has been verified:

  1. feishu_im_user_message.send to a topic-group chat_id creates a new top-level topic.
  2. feishu_im_user_message.reply with reply_in_thread=true posts inside that topic.
  3. A reply can @ a user by placing Feishu text markup directly in the content:
<at user_id="ou_xxx">Name</at> 你好

This is the current working path. Prefer it over assumptions about implicit reply routing.

Preconditions

Require all of the following:

  • The destination is a real Feishu topic-group.
  • The user has explicitly asked to send the message.
  • The user has permission to post in that group.
  • User-auth Feishu messaging is available for feishu_im_user_message.

If user auth is unavailable or expired, stop and ask the user to re-authorize.

If the target group is not a topic-group, explain that the result will only be a normal group message, not a topic.

Default target group

Default to the fixed topic-group configured in TOOLS.md, if one exists.

Do not hardcode a personal/local group name into a public skill.

Use feishu_chat with action=search to resolve the group.

If the user explicitly names a target group, use that group.

If the user does not name a group and no local default is configured:

  • Stop.
  • Ask which Feishu topic-group to use.

If no matching group is found:

  • Stop.
  • Tell the user the topic-group does not exist yet.
  • Ask them to create it manually first.

Workflow

1. Resolve target group

  • If the user explicitly names a group, use that name.
  • Otherwise use the local default topic-group from TOOLS.md, if configured.
  • If no local default exists, ask the user which topic-group to use.
  • Resolve the chat_id with feishu_chat.search.

2. Parse requested action

Support four shapes:

A. Open topic only

Example:

开话题:OpenClaw 多线程 SOP

Send one top-level text message only.

B. Open topic + first thread reply

Example:

开话题:OpenClaw 多线程 SOP|先写一个 5 点大纲

Interpret as:

  • title = OpenClaw 多线程 SOP
  • first thread reply = 先写一个 5 点大纲

C. Open topic + first thread reply + @mention

Example:

开话题:测试 1|回一条:@我 测试

Interpret as:

  • top-level topic title/text
  • then a thread reply
  • optionally prepend a Feishu <at user_id="...">...</at> mention

D. Open topic + carry over prior context + ask a new question

Example:

开个新话题聊这个,把前面几条带过去,然后追问:自由现金流有几个指数?有什么区别?

Interpret as:

  • create a clean new topic
  • carry over only the minimum relevant context
  • put both context and the new question into the same top-level topic-opening message
  • make the new question impossible to miss
  • avoid a second seed message unless the user explicitly wants an extra in-thread reply

3. Create the topic

Use feishu_im_user_message.send:

  • action=send
  • receive_id_type=chat_id
  • receive_id=<chat_id>
  • msg_type=text
  • content={"text":"<top-level text>"}

The returned message_id is the topic root.

4. Optionally add the first thread reply

If the user asked for an initial reply, use feishu_im_user_message.reply:

  • action=reply
  • message_id=<topic root message_id>
  • reply_in_thread=true
  • msg_type=text
  • content={"text":"<reply text>"}

If an @mention is requested, format it directly in the text:

<at user_id="ou_xxx">Name</at> 回复内容

Use the current user open_id when the user says @我.

5. When carrying context into a new topic, keep it simple

For context carry-over + new question, prefer one single top-level message. Do not add a second seed reply unless the user explicitly wants one, because two user messages may trigger two assistant replies.

Use a simple structure:

<标题>

问题:...

前情提要:...

Or, when a short summary reads better:

<标题>

问题:...

前情提要:
- ...
- ...

Rules:

  • Put the real question first.
  • Keep the summary short.
  • Do not add extra instruction-y wording unless absolutely needed.
  • The goal is not to outsmart the model; the goal is to make the follow-up obvious.

Example:

HALO:自由现金流指数有什么区别

问题:
A股里自由现金流目前有几个主流指数?它们分别有什么区别?

前情提要:
- 刚讨论过 HALO 在 A 股的映射,涉及资源、能源、公用事业、电网、交运、央企红利、现金流。
- 这次只是在这个基础上继续追问自由现金流指数。

Parsing guidance

Treat all of the following as likely triggers:

  • 开话题:标题
  • 开个话题:标题
  • 开个新话题:标题
  • /topic 标题
  • /topic 标题|正文
  • 新开一个话题,然后在话题里回复一条
  • 开话题: 标题, 然后在话题里回复 1 条测试并且 at 我

Useful parsing rules:

  • Split on the first / : to isolate the command from the payload.
  • Treat / | as a likely separator between title and first reply.
  • Phrases like 然后回复一条, 回一条, 在话题里回复, 并且 at 我, @我 indicate an in-thread follow-up.
  • Phrases like 把前面几条带过去, 把刚才聊的内容带过去, 新开一个话题聊这个, 继续追问, 新的追问 indicate a context-carrying seed reply rather than a plain freeform reply.
  • If only one segment exists, treat it as topic text only.

Do not overfit parsing. If the message is ambiguous, ask one short clarification question.

Tool preference

Prefer the validated user-message path for this skill:

  • feishu_im_user_message.send
  • feishu_im_user_message.reply

Reason: this path has been validated end-to-end for:

  • topic creation
  • in-thread reply
  • @mention markup

Do not rely on ordinary assistant reply routing to create a fresh topic.

Report back

After success, tell the user briefly:

  • target group
  • topic text/title used
  • whether a thread reply was added
  • whether an @mention was included

Keep it short.

Boundaries

  • This skill only works as intended in a Feishu topic-group.
  • This skill does not create Feishu groups.
  • This skill should not silently send as the user unless the user explicitly asked.
  • This skill is for topic bootstrapping: create the topic, optionally seed the first thread reply, then let later conversation continue inside that topic.
  • If the user wants a normal group message instead of a topic, do not force this skill.

Minimal examples

Example 1: topic only

User:

开话题:测试 4

Action:

  • resolve the configured default topic-group (or ask the user which group to use)
  • send top-level 测试 4

Example 2: topic + thread reply

User:

开话题:测试 3|不 @ 的线程回复测试

Action:

  • create top-level 测试 3
  • reply in thread with 不 @ 的线程回复测试

Example 3: topic + thread reply + @me

User:

开话题:测试 2|@我 再测一次

Action:

  • create top-level 测试 2
  • reply in thread with:
<at user_id="ou_xxx">Name</at> 再测一次

Example 4: carry context + ask one new question cleanly

User:

开个新话题聊这个,把前面几条带过去,然后追问:自由现金流有几个指数?有什么区别?

Action:

  • create a fresh topic title focused on the new question
  • send one top-level topic-opening message
  • put 问题:... first
  • then add a short 前情提要:...
  • do not add a second seed reply unless the user explicitly asks for one

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

73.38%
按下载量换算2,346

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 openclaw skills install feishu-new-chat 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills