ワンクリックでManusで任意のスキルを実行

始める

web-search-provider

スター14

フォーク3

更新日2026年4月17日 13:29

Search the web and extract content from URLs

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

Manusで実行

ソース

gabogabucho

gabogabucho/lumen-agent

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

Manusで実行

Web Search

You can now search the web and extract content from URLs.

How to handle search requests

When a user asks to:

"Search for the latest news about AI"
"What's the weather in Buenos Aires?"
"Find information about Python 3.13"
"Look up the best restaurants in Madrid"

Do the following:

Identify the search query from the user's message
Use web__search with the query
Summarize the results in a clear, concise way
If the user wants more detail, use web__extract on a specific URL

Best practices

Keep summaries concise — 2-3 paragraphs max
Cite sources when providing factual information
If search returns no results, suggest refining the query
For time-sensitive info (weather, news), note that results may not be real-time

Limitations

Web search depends on the web connector handler being configured. For full functionality, consider connecting an MCP server like @anthropic/mcp-server-fetch or using a search API (Tavily, Brave).

このリポジトリの他の Skills

同じリポジトリ

scheduler

gabogabucho/lumen-agent

Schedule reminders and recurring tasks that fire automatically. Use scheduler__create to save them.

2026-06-1014

honcho-memory

gabogabucho/lumen-agent

Integración con Honcho.dev para memoria persistente cross-session — recuperar hechos, aprender de interacciones pasadas, modelado de pares y razonamiento dialéctico.

2026-05-0914

paperclip

gabogabucho/lumen-agent

Integración con Paperclip para orquestación multi-agente. Recibe tareas de Paperclip, procesa con el brain de Lumen y reporta estado al CEO.

2026-05-0914

x-lumen-comunicacion-telegram

gabogabucho/lumen-agent

Tenés Telegram disponible como canal de salida y polling de entrada.

2026-04-2114

file-manager

gabogabucho/lumen-agent

Read, write, and list local files

2026-04-1714

text-responder

gabogabucho/lumen-agent

Responds to questions using the configured LLM

2026-04-1614

name	web-search-provider
description	Search the web and extract content from URLs
min_capability	tier-2
requires	{"connectors":["web"]}

Web Search

You can now search the web and extract content from URLs.

How to handle search requests

When a user asks to:

"Search for the latest news about AI"
"What's the weather in Buenos Aires?"
"Find information about Python 3.13"
"Look up the best restaurants in Madrid"

Do the following:

Identify the search query from the user's message
Use web__search with the query
Summarize the results in a clear, concise way
If the user wants more detail, use web__extract on a specific URL

Best practices

Keep summaries concise — 2-3 paragraphs max
Cite sources when providing factual information
If search returns no results, suggest refining the query
For time-sensitive info (weather, news), note that results may not be real-time

Limitations

Web search depends on the web connector handler being configured. For full functionality, consider connecting an MCP server like @anthropic/mcp-server-fetch or using a search API (Tavily, Brave).