← Back to Skills Marketplace
chaoliuzhu

德胧舆情情报综合工具箱

by Chaoliuzhu · GitHub ↗ · v2.0.1 · MIT-0
cross-platform ⚠ suspicious
123
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install delonix-intelligence-suite-v2
Description
德胧AI龙虾军团舆情情报综合工具箱 v2.0。整合 miaoda搜索 + BettaFish开源舆情引擎 + 飞书精美卡片 + 抖音/视频平台监控 + AI风险分析与防控预案生成。 触发词:舆情、情报、网情、监控、风险分析、抖音监控、酒店风险、德胧舆情、竞品动态、行业预警 三大核心能力: 1. 外网信息采集(mi...
README (SKILL.md)

德胧舆情情报综合工具箱 v2.0

德胧AI龙虾军团专属舆情情报系统 版本:v2.0 | 日期:2026-04-19 | 作者:晁留柱的助手(小柱) 整合:miaoda-web-search + BettaFish + 飞书卡片 + AI风险分析


🔧 核心架构

德胧舆情情报综合工具箱 v2.0
├── 数据源层
│   ├── miaoda-web-search(主力)→ 关键词搜索 + AI摘要
│   ├── BettaFish扩展 → 30+平台深度采集(微博/小红书/抖音/快手/知乎)
│   └── 视频平台专项 → 抖音/快手/B站 视频舆情
├── 处理层
│   ├── 关键词模板(华住/锦江/亚朵/德胧/首旅如家)
│   ├── 视频舆情过滤(抖音含"德胧"关键字)
│   ├── 热度排序(按发布时间/媒体权重)
│   └── 去重+排重
├── AI分析层
│   ├── 舆情分类(投诉/安全/服务/运营/AI)
│   ├── 风险评级(高危/中危/低危)
│   └── 防控预案生成(针对同类舆情)
├── 输出层
│   ├── 飞书精美卡片(日报/突发/预警)
│   └── 结构化数据表 → 入库分析
└── 自动化层
    └── cron每日9:00自动采集推送

⚡ 快速使用

舆情采集(基础)

# 酒店舆情搜索
miaoda-studio-cli search-summary --query "华住 回访电话 隐私泄露 2026年4月"

# 抖音/视频平台舆情
miaoda-studio-cli search-summary --query "抖音 德胧 开元名都 投诉" --instruction "重点关注视频内容,提炼视频核心观点"

# 竞品动态
miaoda-studio-cli search-summary --query "锦江 AI动态定价 2026" --instruction "提取AI覆盖数据、效果数据、竞品对比"

风险分析(AI驱动)

触发词:风险分析、防控预案、规避建议

输入:舆情数据(投诉内容/事件描述)
输出:
1. 风险评级(高危/中危/低危)
2. 风险归因(根因分析)
3. 防控预案(具体措施)
4. 预警指标(监控阈值)

🛠️ 工具一:miaoda-web-search(主力·已集成)

简介

妙搭内置网页搜索工具,通过 miaoda-studio-cli search-summary 实现。

命令格式

miaoda-studio-cli search-summary --query "关键词" [--instruction "AI处理指令"] [--output text|json]

参数说明

参数 说明 必需
--query 搜索关键词(1-500字符)
--instruction AI处理指令
--output 输出格式:text/json

🛠️ 工具二:BettaFish 微舆系统(扩展)

简介

BettaFish(微舆) 是GitHub上最强的开源舆情分析系统(666ghj/BettaFish),支持30+平台深度采集。

仓库信息

GitHub: https://github.com/666ghj/BettaFish
Star: 17k+
语言: Python3
架构: 多Agent + LLM + NLP

核心能力

能力 说明
平台覆盖 微博/小红书/抖音/快手/B站/知乎/公众号等30+
AI分析 情感分析/话题聚类/传播路径/趋势预测
多模态 突破图文限制,深度解析短视频内容
Agent协作 多Agent辩论机制,链式思维碰撞
报告生成 自动生成专业舆情分析报告

安装状态

⚠️ 待安装 - 受网络限制,pip/GitHub下载可能超时。

安装步骤(网络恢复后)

# 方式1:pip安装
pip install bettafish

# 方式2:源码克隆
git clone [email protected]:666ghj/BettaFish.git /tmp/BettaFish
cd /tmp/BettaFish
pip install -r requirements.txt

# 方式3:Docker(推荐)
docker pull bettafish/bettafish:latest
docker run -d bettafish/bettafish

使用示例

# 基础舆情分析
bettafish analyze --keyword "华住 酒店 投诉" --platforms weibo,xiaohongshu,douyin

# 视频平台专项
bettafish analyze --keyword "德胧" --platforms douyin,kuaishou,bilibili --video-only

# 风险报告生成
bettafish report --type risk --output json

# 竞品对比分析
bettafish compare --brands 华住,锦江,亚朵 --metric reputation

BettaFish + miaoda 协同方案

阶段1(当前):miaoda快速搜索
└── 负责:关键词搜索、AI摘要、实时热点发现

阶段2(扩展):BettaFish深度分析
└── 负责:30+平台结构化采集、情感分析、风险评级、视频舆情

协同流程:
用户请求 → miaoda快速发现 → BettaFish深度分析 → AI风险评估 → 飞书卡片输出

🛠️ 工具三:抖音/视频平台专项监控

视频舆情特点

特点 说明
传播速度快 视频发酵速度远超图文
匿名性强 博主身份难追溯
情绪化 视觉冲击更强,负面情绪放大
二次创作 易引发模仿和扩散

抖音舆情关键词模板

# 德胧品牌词
miaoda-studio-cli search-summary --query "抖音 开元名都 德胧 投诉"
miaoda-studio-cli search-summary --query "抖音 钛唐酒店 德胧 差评"
miaoda-studio-cli search-summary --query "抖音 德胧集团 酒店 卫生"

# 竞品酒店视频舆情
miaoda-studio-cli search-summary --query "抖音 华住 酒店 隐私 泄露"
miaoda-studio-cli search-summary --query "抖音 桔子水晶 酒店 投诉"

视频舆情分级

级别 特征 响应时间
P0红色 视频播放量>10万、负面评论>1000 30分钟内
P1橙色 播放量>1万、负面评论>100 2小时内
P2黄色 播放量>1000、有扩散趋势 4小时内
P3蓝色 普通投诉视频 24小时内

🛠️ 工具四:飞书精美卡片生成器

飞书卡片模板 v2.0

基于 ritaswc/lark-card-message-builder 和飞书官方Card JSON 2.0结构。

精美卡片JSON结构

{
  "config": {"wide_screen_mode": true},
  "header": {
    "template": "red",
    "title": {"content": "🔥 酒店舆情日报 | 2026-04-19", "tag": "plain_text"}
  },
  "elements": [
    {"tag": "markdown", "content": "**\x3Ctext_tag color='red'>🔴 超级热点\x3C/text_tag> [标题]**\
\
[内容]"},
    {"tag": "hr"},
    {"tag": "markdown", "content": "**\x3Ctext_tag color='orange'>🟠 投诉数据\x3C/text_tag>**\
\
• [投诉1]\
• [投诉2]"},
    {"tag": "hr"},
    {"tag": "markdown", "content": "**\x3Ctext_tag color='green'>🟡 风险分析\x3C/text_tag>**\
\
[AI分析内容]"},
    {"tag": "note", "content": "🦞 德胧AI龙虾军团 | v2.0", "quoteless": false}
  ]
}

飞书Markdown语法要点

要素 语法 注意
标题 **加粗文字** ❌ 不支持 # ## ###
标签 \x3Ctext_tag color='red'> 颜色:red/orange/yellow/green/blue/purple
链接 [文本](url) -
分割线 --- -
列表 • 项目1. 项目 bullet list更稳定
彩色文本 \x3Cfont color='red'>文字\x3C/font> -

发送精美卡片

# 使用 feishu_im_user_message 工具
# msg_type: interactive
# content: JSON字符串(见上方模板)

🛠️ 工具五:AI风险分析与防控预案

风险分析流程

舆情数据输入
    ↓
① 舆情分类(5类)
    ├─ 投诉类(服务/卫生/退改)
    ├─ 安全类(隐私/消防/事故)
    ├─ 服务类(态度/效率/承诺)
    ├─ 运营类(加盟/关店/业绩)
    └─ AI类(定价/隐私/自动化)
    ↓
② 风险评级
    ├─ 高危(媒体曝光>5家/投诉量>100)
    ├─ 中危(媒体曝光>2家/投诉量>30)
    └─ 低危(个别投诉/无媒体曝光)
    ↓
③ 根因分析(5Why)
    ↓
④ 防控预案生成
    ├─ 预防措施
    ├─ 监控指标
    ├─ 应急预案
    └─ 责任人
    ↓
⑤ 输出飞书卡片

风险分析Prompt模板

请分析以下舆情事件,生成风险报告:

## 舆情内容
[粘贴舆情内容]

## 分析要求
1. 风险评级(高危/中危/低危)
2. 风险归因(根因分析,5Why法)
3. 防控预案:
   - 预防措施(具体可执行)
   - 监控指标(可量化)
   - 应急预案(触发条件+处置流程)
4. 同类舆情规避建议

## 输出格式
请用飞书Markdown格式输出,包含:
- 风险评级(高危用🔴标记)
- 根因分析
- 防控预案(表格形式)
- 规避建议

防控预案示例

类别 预防措施 监控指标 应急预案
隐私泄露 回访电话需用户授权、话术标准化 投诉量>5/月 立即下架回访、隐私合规培训
卫生投诉 客房SOP增加抽查频次 差评率>2% 门店深度清洁、公开道歉
强制好评 禁止任何形式强制、投诉零容忍 投诉量>3/月 涉事员工追责、品牌公开澄清

📋 德胧专属关键词库 v2.0

企业监控(扩展)

企业 品牌词 投诉关键词
华住会 华住、桔子、汉庭、全季、海友、漫心、美居 隐私、泄露、好评、威胁
锦江集团 锦江、锦江酒店、锦鲲 定价、加盟、解约
亚朵 亚朵、Atour 服务、卫生、退改
首旅如家 首旅如家、如家、和颐 卫生、安全、投诉
德胧 开元名都、钛唐、开元酒店 [德胧专属]

视频平台关键词

平台 搜索关键词 重点关注
抖音 开元名都+投诉、德胧+曝光、钛唐+差评 视频评论、博主观点
快手 酒店+投诉、德胧+事件 老铁文化、传播速度
B站 酒店行业、隐私泄露、酒店评测 UP主深度测评
小红书 酒店体验、投诉笔记、酒店避雷 素人笔记、真实体验

事件类型关键词

类型 关键词组合
隐私安全 隐私泄露、信息泄露、摄像头、入住记录
服务质量 卫生差、前台态度、虚假宣传、货不对板
会员权益 积分清零、会员降级、权益缩水、取消优惠
价格问题 临时涨价、差价、取消扣款、价格欺诈
安全事故 消防隐患、安全事故、食物中毒、意外伤害

⏰ 自动化配置

Cron:每日舆情日报

{
  "name": "德胧舆情日报v2.0",
  "schedule": {
    "kind": "cron",
    "expr": "0 9 * * *",
    "tz": "Asia/Shanghai"
  },
  "sessionTarget": "isolated",
  "payload": {
    "kind": "agentTurn",
    "message": "执行德胧舆情日报v2.0采集:\
1. miaoda搜索今日最新舆情(华住/锦江/亚朵/德胧)\
2. 抖音/视频平台专项搜索(抖音+快手+B站)\
3. AI风险分析(高危舆情生成防控预案)\
4. 生成飞书精美卡片(日报格式)\
5. 发送私信到 ou_fc1e75d64fec6e10ce94a51adc6f6409"
  },
  "delivery": {
    "mode": "announce",
    "channel": "feishu",
    "to": "ou_fc1e75d64fec6e10ce94a51adc6f6409"
  }
}

Cron:突发舆情预警

{
  "name": "德胧突发舆情预警",
  "schedule": {
    "kind": "cron",
    "expr": "0 */2 * * *",
    "tz": "Asia/Shanghai"
  },
  "sessionTarget": "isolated",
  "payload": {
    "kind": "agentTurn",
    "message": "执行突发舆情检查:\
1. 搜索最新抖音/视频平台舆情(含'德胧'关键字)\
2. 检查P0/P1级风险\
3. 如有突发高危舆情,立即生成预警卡片发送私信"
  },
  "delivery": {
    "mode": "announce",
    "channel": "feishu",
    "to": "ou_fc1e75d64fec6e10ce94a51adc6f6409"
  }
}

🔄 迭代计划

v2.0(当前)

  • miaoda-web-search 集成
  • BettaFish 扩展指南
  • 抖音/视频平台专项监控
  • 飞书精美卡片模板 v2.0
  • AI风险分析Prompt模板
  • BettaFish 本地部署(待网络恢复)
  • 防控预案数据库

v3.0(计划中)

  • BettaFish 本地化部署
  • 抖音API官方接入(需企业认证)
  • 可视化舆情仪表盘
  • 微信视频号监控

v4.0(长期)

  • 实时预警Webhook推送
  • 竞品对比分析报告
  • 季度舆情白皮书

🦞 部署指南

一键安装(本Skill)

# 1. 复制Skill到OpenClaw目录
cp -r delonix-intelligence-suite-v2/ ~/workspace/agent/skills/

# 2. 重启OpenClaw
openclaw gateway restart

# 3. 验证
miaoda-studio-cli search-summary --query "测试" --instruction "验证工具是否正常"

BettaFish安装(网络恢复后)

# 克隆BettaFish
git clone [email protected]:666ghj/BettaFish.git /tmp/BettaFish

# 安装依赖
cd /tmp/BettaFish
pip install -r requirements.txt

# 配置API Key(用于LLM分析)
export OPENAI_API_KEY="your-key"
export BETTAFISH_CONFIG="/path/to/config.yaml"

⚠️ 注意事项

  1. 网络限制:GitHub HTTPS可能超时,优先使用SSH克隆
  2. 数据合规:采集内容仅用于内部分析,不可公开传播
  3. 频率控制:搜索请求间隔建议>3秒
  4. 视频舆情:抖音等平台可能需要Cookie认证
  5. 内容审核:AI分析结果需人工核实

📞 技术支持

  • SKILL作者:晁留柱的助手(小柱)
  • 版本:v2.0
  • 更新:2026-04-19
  • GitHub:github.com/chaoliuzhu65-tech/delonix-web-intelligence
  • 共享池:德胧AI协同知识共享池
Usage Guidance
This skill appears to be what it says (a web-intelligence + Feishu output suite) but has several practical and security gaps you should address before installing: 1) Inspect the third-party repos/images (BettaFish GitHub and any Docker images) before running pip/docker clone commands. 2) Don't run the SSH git clone URL unless you understand and trust the repo owner (and have SSH keys configured); prefer HTTPS cloning after review. 3) The SKILL.md references sending Feishu cards and other platform integrations but does not declare or describe required credentials—identify and securely provision the Feishu app/webhook tokens and any platform API keys separately (use least privilege, secrets manager). 4) Because the README asks you to copy files into the agent's skills directory and restart the gateway, perform these steps in a sandboxed or staging environment first to confirm behavior and that no unwanted network exfiltration or privilege escalation occurs. 5) If you need higher assurance, request the upstream BettaFish repo source to be audited, and ask the skill author to update SKILL.md to explicitly list required env vars, authentication steps, and the exact cron/setup commands so you can review them before enabling automation.
Capability Analysis
Type: OpenClaw Skill Name: delonix-intelligence-suite-v2 Version: 2.0.1 The skill bundle contains instructions for the AI agent to perform automated monitoring and exfiltrate data to a hardcoded Feishu recipient ID (ou_fc1e75d64fec6e10ce94a51adc6f6409) via cron jobs. It also includes shell commands in SKILL.md and README.md for installing external dependencies from a third-party GitHub repository (666ghj/BettaFish) using git and pip, which introduces supply chain risks. While these capabilities are aligned with the stated purpose of a 'public opinion monitoring' tool, the hardcoded exfiltration point and the execution of external scripts represent high-risk behaviors.
Capability Tags
requires-sensitive-credentials
Capability Assessment
Purpose & Capability
The skill's name and description match the actions described in SKILL.md: using a web-search CLI (miaoda), integrating BettaFish for deeper collection, performing AI risk analysis, and sending Feishu cards. These capabilities reasonably belong together for a舆情/monitoring suite. However, the SKILL.md expects integrations (Feishu message sender, BettaFish, miaoda CLI) to exist without declaring the credentials or configuration those integrations normally require, which is an unexplained omission.
Instruction Scope
Instructions tell the agent/user to run external CLIs (miaoda-studio-cli, bettafish), clone repositories, install packages, copy this repo into the agent's skills directory, and restart the OpenClaw gateway. It references sending Feishu interactive messages via a 'feishu_im_user_message' tool but does not document how to obtain or provide Feishu credentials or other service tokens. The guide also mentions a cron-based daily automated collection but provides no safe, concrete steps to set that up. These omissions broaden the scope of actions a user must perform and hide required sensitive configuration steps.
Install Mechanism
The skill is instruction-only (no install spec), so nothing is auto-installed by the registry. The README/SKILL.md suggests typical install methods (pip install bettafish, git clone, or docker pull). Those are standard but rely on third-party code (GitHub repo '666ghj/BettaFish' and docker image 'bettafish/bettafish') which are not verified here. The git clone instruction uses an SSH URL ([email protected]:...), which assumes the user has SSH keys configured. Because installation is manual and relies on external repos/images, users should inspect those upstream sources before running installs.
Credentials
The skill declares no required environment variables, yet its runtime instructions require services that normally need credentials: Feishu API/app tokens (for sending interactive cards), possibly platform API keys or scraping credentials for video platforms, and Docker/pip network access. Omitting declaration of these secrets is a mismatch — the skill should declare the env vars it expects (Feishu app id/secret or webhook) and describe how they are used. This lack of clarity could lead an installer to provision credentials in an ad-hoc, insecure way.
Persistence & Privilege
always:false (normal). The README instructs copying the repo into the agent's skills directory and restarting the OpenClaw gateway, which is expected for installing a skill but does grant filesystem modification and persistent presence. The guide also describes a cron-based daily automated collection (no code supplied). This combination means the skill may be run on a schedule once installed; users should be aware that installation implies ongoing activity but the skill itself does not request 'always' or autonomic privileges in the registry metadata.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install delonix-intelligence-suite-v2
  3. After installation, invoke the skill by name or use /delonix-intelligence-suite-v2
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.0.1
v2.0.1优化舆情分析逻辑,新增风险预警
v2.0.0
整合舆情+飞书卡片+AI分析
Metadata
Slug delonix-intelligence-suite-v2
Version 2.0.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is 德胧舆情情报综合工具箱?

德胧AI龙虾军团舆情情报综合工具箱 v2.0。整合 miaoda搜索 + BettaFish开源舆情引擎 + 飞书精美卡片 + 抖音/视频平台监控 + AI风险分析与防控预案生成。 触发词:舆情、情报、网情、监控、风险分析、抖音监控、酒店风险、德胧舆情、竞品动态、行业预警 三大核心能力: 1. 外网信息采集(mi... It is an AI Agent Skill for Claude Code / OpenClaw, with 123 downloads so far.

How do I install 德胧舆情情报综合工具箱?

Run "/install delonix-intelligence-suite-v2" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is 德胧舆情情报综合工具箱 free?

Yes, 德胧舆情情报综合工具箱 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does 德胧舆情情报综合工具箱 support?

德胧舆情情报综合工具箱 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created 德胧舆情情报综合工具箱?

It is built and maintained by Chaoliuzhu (@chaoliuzhu); the current version is v2.0.1.

💬 Comments