web-search

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Web Search Skill

网页搜索Skill

Search the web. Prefer the built-in WebSearch tool — it uses a real browser engine with high success rate. Fall back to curl only if WebSearch is unavailable.

进行网页搜索。优先使用内置的WebSearch工具——它使用真实的浏览器引擎，成功率更高。仅当WebSearch不可用时才退而使用curl。

DuckDuckGo Instant Answer API (No API key needed)

DuckDuckGo即时答案API（无需API密钥）

bash

undefined

bash

undefined

Basic search — returns instant answers

curl -s "https://api.duckduckgo.com/?q=QUERY&format=json&no_html=1" | jq '.Abstract, .RelatedTopics[:5]'

undefined

curl -s "https://api.duckduckgo.com/?q=QUERY&format=json&no_html=1" | jq '.Abstract, .RelatedTopics[:5]'

undefined

DuckDuckGo HTML Search (Full results)

DuckDuckGo HTML搜索（完整结果）

bash

undefined

bash

undefined

Get search results page and extract links

curl -s "https://html.duckduckgo.com/html/?q=QUERY" | grep -o 'href="https{0,1}://[^"]*' | head -10

undefined

curl -s "https://html.duckduckgo.com/html/?q=QUERY" | grep -o 'href="https{0,1}://[^"]*' | head -10

undefined

Google Search via SerpAPI (if API key available)

通过SerpAPI进行Google搜索（需API密钥）

bash

undefined

bash

undefined

Check if SERPAPI_KEY is set

alma config get serpapi.apiKey

If available:

curl -s "https://serpapi.com/search.json?q=QUERY&api_key=API_KEY" | jq '.organic_results[:5] | .[] | {title, link, snippet}'

undefined

curl -s "https://serpapi.com/search.json?q=QUERY&api_key=API_KEY" | jq '.organic_results[:5] | .[] | {title, link, snippet}'

undefined

Fetching Page Content

获取页面内容

After finding URLs from search, use the WebFetch tool to get the actual page content:

WebFetch(url="https://example.com/article")

Or via curl:

bash

curl -sL "https://example.com/article" | head -200

从搜索结果中找到URL后，使用WebFetch工具获取实际页面内容：

WebFetch(url="https://example.com/article")

或者通过curl:

bash

curl -sL "https://example.com/article" | head -200

Tips

提示

URL-encode the query: replace spaces with
```
+
```
or
```
%20
```
Use
```
jq
```
to parse JSON responses
For complex queries, try multiple search approaches
Always summarize findings for the user rather than dumping raw results
If DuckDuckGo doesn't have good results, try fetching specific known sources directly

对查询内容进行URL编码：将空格替换为
```
+
```
或
```
%20
```
使用
```
jq
```
解析JSON响应
对于复杂查询，尝试多种搜索方式
始终为用户总结搜索结果，而非直接输出原始内容
如果DuckDuckGo的结果不理想，尝试直接获取特定的已知来源内容

Examples

示例

"最近的 AI 新闻":

bash

curl -s "https://html.duckduckgo.com/html/?q=latest+AI+news+2026" | grep -o 'href="https\{0,1\}://[^"]*' | grep -v duckduckgo | head -5

"Python 3.13 新特性":

bash

curl -s "https://api.duckduckgo.com/?q=python+3.13+new+features&format=json&no_html=1" | jq '.Abstract'

"最近的 AI 新闻":

bash

curl -s "https://html.duckduckgo.com/html/?q=latest+AI+news+2026" | grep -o 'href="https\{0,1\}://[^"]*' | grep -v duckduckgo | head -5

"Python 3.13 新特性":

bash

curl -s "https://api.duckduckgo.com/?q=python+3.13+new+features&format=json&no_html=1" | jq '.Abstract'