← Back to Skills Marketplace
Rumour Buster
by
harry720320
· GitHub ↗
· v0.3.0
· MIT-0
87
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install rumour-buster
Description
谣言终结者。双引擎验证消息真假：中文搜索（kimi_search + multi-search-engine）+ 英文搜索（Tavily + multi-search-engine），交叉验证，溯源消息来源。首次使用自动运行 setup 配置。
Usage Guidance
要点与建议：
- 该技能会在首次运行时自动进入交互式 setup 并在 ~（家目录）写入 ~/.rumour-buster-config，确认你愿意在本机保存该配置文件。
- scripts/tavily_search.py 包含一个硬编码的 Tavily 开发 API key；如果你不提供自己的 key，查询将使用该内置 key 并把查询发送到 https://api.tavily.com，这可能导致隐私泄露或把你的查询记录到第三方。建议在环境中设置你自己的 TAVILY_API_KEY 或从代码中移除/替换默认键。
- 如果你担心数据外泄或凭据滥用：在安装前审阅 sub-skills/setup/setup.py 和 scripts/tavily_search.py，或在隔离环境里运行；也可以删除/覆盖硬编码默认 key 并只使用你自己的 key。
- 该技能要求已安装 api-multi-search-engine（通过路径检测），会建议用 clawhub 安装，这是与功能相符的依赖。
- 结论：功能上合理且与描述一致，但嵌入的默认 API key 和自动写配置的实现增加了隐私/凭据风险，安装前请评估并采取上面的缓解措施。
Capability Analysis
Type: OpenClaw Skill
Name: rumour-buster
Version: 0.3.0
The skill bundle contains a hardcoded Tavily API key in `scripts/tavily_search.py` (tvly-dev-3QAy8g-...), which is a security vulnerability (credential exposure) and could allow the key owner to monitor user search queries. Furthermore, there is a functional inconsistency: while `sub-skills/setup/setup.py` prompts the user for their own API key and saves it to a hidden configuration file (`~/.rumour-buster-config`), the search script `tavily_search.py` does not read this configuration file, potentially causing the agent to use the hardcoded key or fail to use the user's configured credentials unless environment variables are manually set.
Capability Assessment
Purpose & Capability
技能名与描述（中文+英文多引擎交叉验证）与包含的脚本和 setup 流程一致：有中文/英文搜索逻辑、setup 子技能用于检测并写入 ~/.rumour-buster-config、以及一个用于调用 Tavily 的脚本。总体能力与目的对齐，但 Tavily 集成会向外部 API 发送用户查询，这是功能上预期的。
Instruction Scope
SKILL.md 明确指示会自动运行 setup（首次使用）并生成 ~/.rumour-buster-config；setup.py 会与用户交互请求/接收 Tavily API Key 并把配置写入家目录。说明没有要求读取系统上与功能无关的文件或凭据，但自动运行 setup 和写配置是需要用户知晓的行为。
Install Mechanism
这是一个 instruction-first 技能（无安装规范），源码仅包含两个 Python 脚本和文档，没有远程下载或可执行安装步骤，风险较低。
Credentials
元数据不声明任何必需环境变量，但 scripts/tavily_search.py 使用 os.getenv('TAVILY_API_KEY') 并包含一个硬编码的默认 API key (tvly-dev-...)：如果用户未提供自己的 key，查询将使用这个嵌入 key 发往 api.tavily.com，带来隐私与凭据滥用风险。技能还引用（检测）~/.openclaw/workspace/skills 路径以判断依赖，写入 ~/.rumour-buster-config。总体请求的凭据与技能目的相关，但硬编码 key 与未声明的外部数据流是不成比例的风险点。
Persistence & Privilege
技能不会声明 always:true，也不会修改其他技能的配置；唯一的持久化是写入其自己的配置文件 (~/.rumour-buster-config)，这是与其功能一致的常见行为。
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install rumour-buster - After installation, invoke the skill by name or use
/rumour-buster - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.3.0
**Major update: Adds automatic setup, cross-language source tracing, and enhanced dual-engine fact-checking.**
- Introduced automatic setup on first use, guiding users to configure search engines and optional Tavily API.
- Enhanced verification by combining Chinese (kimi_search + multi-search-engine) and English (Tavily + multi-search-engine) searches.
- Added cross-language source tracing, earliest source detection, and path tracking for forwarded information.
- Output now features structured verification reports with trust scores, sources, and step-by-step findings.
- Manual (re)setup supported via setup commands.
Metadata
Frequently Asked Questions
What is Rumour Buster?
谣言终结者。双引擎验证消息真假：中文搜索（kimi_search + multi-search-engine）+ 英文搜索（Tavily + multi-search-engine），交叉验证，溯源消息来源。首次使用自动运行 setup 配置。 It is an AI Agent Skill for Claude Code / OpenClaw, with 87 downloads so far.
How do I install Rumour Buster?
Run "/install rumour-buster" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Rumour Buster free?
Yes, Rumour Buster is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Rumour Buster support?
Rumour Buster is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Rumour Buster?
It is built and maintained by harry720320 (@harry720320); the current version is v0.3.0.
More Skills