← 返回 Skills 市场
107
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install hot-search
功能描述
Hot Search - OpenClaw 稳定搜索技能,专为金融数据和市场行情设计。支持多引擎聚合搜索,无需 API 密钥,免费无限次使用。
使用说明 (SKILL.md)
Hot Search 搜索技能
版本:1.0.0 | 作者:糖糖 (Tangtang) | 邮箱:[email protected]
🔍 技能概述
Hot Search 是专为 OpenClaw 框架设计的稳定搜索技能,特别优化了金融数据和市场行情的搜索能力。
核心优势
- 稳定可靠 - 多引擎备份,单引擎失败不影响全局
- 免费使用 - 无需 API 密钥,无使用次数限制
- 快速响应 - 超时控制,避免长时间阻塞
- 金融优化 - 针对原油、股票等金融数据优化
📦 安装方法
方法 1:从 Clawhub 安装
openclaw skill install hot-search
方法 2:从 GitHub 安装
git clone https://github.com/fishsomes/hot-search.git
cd hot-search
pip3 install -r requirements.txt
方法 3:手动安装
将技能文件复制到 OpenClaw 技能目录:
cp -r hot-search ~/.openclaw/workspace/skills/
🚀 使用指南
命令行使用
# 基本搜索
python3 search_skill.py "关键词"
# 指定引擎
python3 search_skill.py "关键词" bing_global
# 多引擎搜索
python3 search_skill.py "关键词" all
Python 调用
from search_skill import SearchEngine
search = SearchEngine()
# 单引擎搜索
results = search.search('原油价格', 'bing_global')
# 多引擎搜索
results = search.search_all('Oman crude oil price')
# 去重结果
unique_results = search.deduplicate(results)
OpenClaw 集成
在 OpenClaw 中直接使用:
用 hot-search 搜索【原油价格】
用 hot-search 的必应国际搜索【WTI crude oil】
🔧 高级配置
超时设置
search = SearchEngine(timeout=5) # 5 秒超时
延迟控制
search = SearchEngine(delay_range=(1.0, 2.0)) # 1-2 秒延迟
自定义引擎
engines = {
"custom": {
"name": "自定义引擎",
"url": "https://example.com/search",
"params": {"q": "{keyword}"},
"selector": ".result"
}
}
📊 搜索结果格式
{
"bing_cn": [
{
"title": "结果标题",
"link": "https://...",
"snippet": "摘要内容"
}
],
"bing_global": [...],
"yandex": [...],
"swisscows": [...]
}
💡 最佳实践
金融数据搜索
# 使用必应国际搜索英文数据
python3 search_skill.py "DME Oman crude oil settlement price" bing_global
# 使用必应国内搜索中文数据
python3 search_skill.py "中国原油进口量 2026" bing_cn
新闻搜索
# 搜索最新新闻
python3 search_skill.py "今天最新新闻 热点" bing_cn
# 搜索国际新闻
python3 search_skill.py "world news today breaking" bing_global
组合使用
# 1. 先搜索关键词
results = search.search_all('原油价格')
# 2. 筛选优质链接
good_links = [r['link'] for r in results['bing_global'][:3]]
# 3. 深度抓取内容
for link in good_links:
content = web_fetch(url=link)
print(content)
⚠️ 注意事项
- 合法合规 - 仅用于合法数据抓取
- 频率控制 - 避免高频请求被封禁
- 超时设置 - 建议 2-5 秒超时
- 错误处理 - 捕获异常,记录失败
🔗 相关资源
最后更新:2026-03-28 作者:FishSome 邮箱:[email protected]
安全使用建议
This skill appears to do what it says (aggregate web searches and download images). Before installing: 1) Note the script performs web scraping and will make arbitrary outbound HTTP requests and write files (images) to disk — change the default output_dir to a safe path. 2) The README mentions 'Trading Economics' but the code doesn't include that engine; verify the repo if you rely on that source. 3) Downloaded files come from third-party URLs — treat them cautiously (scan or sandbox before opening). 4) Run in a sandbox/container or non-privileged account, and review the code if you plan to run it automatically. 5) Respect target sites' terms of service and rate limits; use delay_range/timeout and proxies responsibly to avoid abuse or IP blocking.
功能分析
Type: OpenClaw Skill
Name: hot-search
Version: 1.0.0
The skill provides multi-engine search and image downloading capabilities but contains a path traversal vulnerability in `search_skill.py`. The `search_and_download` function uses the unsanitized `keyword` input to construct local file paths via `os.path.join`, which could allow an attacker to write files to unauthorized locations if the agent is prompted with a malicious keyword (e.g., containing directory traversal sequences). Additionally, the code includes a hardcoded absolute path (`/home/fishsome/.openclaw/workspace/tmp`) specific to the author's environment, indicating a lack of input validation and portability.
能力评估
Purpose & Capability
The code (search_skill.py) implements multi-engine search (Bing domestic/global, Yandex, Swisscows) and image download, which aligns with the skill name and description. Documentation/README claim of 'integrated Trading Economics' is not reflected in SEARCH_ENGINES; this is a documentation mismatch and should be clarified. The script's behavior (scraping, image download) is consistent with a search/market-data helper.
Instruction Scope
SKILL.md instructs normal installation and usage (git clone, pip install, run python). It does not request credentials or instruct the agent to read unrelated system files. Minor issues: an example uses web_fetch(url=link) which is not provided by the skill (an example helper only); SKILL.md also suggests using proxies/User-Agent which is expected for scraping. The script writes downloaded images to disk (default output_dir set to /home/fishsome/.openclaw/workspace/tmp) — expected for download functionality but users should be aware of filesystem writes.
Install Mechanism
No install spec in the registry (instruction-only), but repository-based installation via pip from requirements.txt is recommended. Dependencies are standard Python packages (requests, beautifulsoup4, lxml) from PyPI — no unusual or remote archive downloads detected.
Credentials
The skill declares no required environment variables, no credentials, and does not access sensitive system configs. It only performs network requests to search engines and to URLs returned by those searches, which is proportionate to its stated function.
Persistence & Privilege
The skill is not set to always:true, does not request persistent privileges, and does not modify other skills or global agent configuration. Autonomous invocation is enabled by default (standard for skills) but is not combined with broad credentials or suspicious behavior here.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install hot-search - 安装完成后,直接呼叫该 Skill 的名称或使用
/hot-search触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Hot Search 1.0.0 - First public release
- Provides stable, multi-engine financial and market data search optimized for OpenClaw.
- No API key required and no usage limit; free and ready to use.
- Supports Bing (CN/global), Yandex, Swisscows, and custom search engines.
- Designed for fast response with timeout and delay configuration.
- Includes detailed usage instructions for command line, Python API, and OpenClaw integration.
元数据
常见问题
Hot Search 是什么?
Hot Search - OpenClaw 稳定搜索技能,专为金融数据和市场行情设计。支持多引擎聚合搜索,无需 API 密钥,免费无限次使用。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 107 次。
如何安装 Hot Search?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install hot-search」即可一键安装,无需额外配置。
Hot Search 是免费的吗?
是的,Hot Search 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Hot Search 支持哪些平台?
Hot Search 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 Hot Search?
由 fishsome(@fishsome)开发并维护,当前版本 v1.0.0。
推荐 Skills