brave-search

Search web, news, images, videos, and local results via Brave Search MCP.

2|Updated Apr 10, 2026
One-click install
npx skills add https://github.com/t0lab/harness-kit --skill brave-search-t0lab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: brave-search
Source: https://github.com/t0lab/harness-kit/tree/main/examples/basic-node/.agents/skills/brave-search
Command: npx skills add https://github.com/t0lab/harness-kit --skill brave-search-t0lab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides concise operational guidance for invoking Brave Search MCP tools to retrieve targeted web, news, image, video, and local search results and to summarize or paginate results, helping agents choose Brave only when the Brave Search MCP is available and appropriate and avoid using it for generic lookups better served by other tools. It clarifies required parameters, safe search settings, and how to obtain and use the BRAVE_API_KEY.

Core Features & Use Cases

  • Tool selection guidance: When to call brave_web_search, brave_news_search, brave_image_search, brave_video_search, brave_local_search, and when to follow with brave_summarizer.
  • Search operators and query tuning: Use operators like site:, ext:, intitle:, lang:, AND/OR/NOT and combine them with freshness, count, safesearch, country, result_filter, and goggles for custom ranking.
  • Summarization and pagination: Request summaries by setting summary=true on brave_web_search then call brave_summarizer with the returned key; check has_more and use offset for pagination.
  • Response shape and fields: Results include title, URL, description, published date, thumbnail, structured data (products, recipes, articles, ratings, FAQs), and a has_more flag for additional pages.
  • Use cases: Finding recent announcements and tutorials, extracting documentation pages from a specific domain, sourcing images or videos for content, and locating businesses or places via local search.

Quick Start

Perform a focused web search for React server components using brave_web_search with freshness set to pm and count 5 and then summarize the top result with brave_summarizer.

Frequently Asked Questions about brave-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I use Brave Search operators for web search queries?

Brave Search operators for web search queries allow you to refine results using site:, ext:, intitle:, and lang: combined with AND/OR/NOT logic to filter documentation, file types, and domains.

Can I summarize Brave Search results and paginate through them?

You can summarize Brave Search results by setting summary=true on brave_web_search, then calling brave_summarizer with the returned key, and paginate using the offset parameter when the has_more flag is true.

Do I need an API key to use Brave Search for news and image queries?

You need a valid BRAVE_API_KEY environment variable to use Brave Search for news and image queries, ensuring the Brave Search MCP is available and properly configured as your preferred search backend.

What fields are included in Brave Search local and video results?

Brave Search local and video results include structured fields such as title, URL, description, published date, thumbnail, structured data like ratings, and a has_more flag for fetching additional pages.

How do I filter Brave Search results by freshness and country?

You can filter Brave Search results by adjusting the freshness parameter for time-sensitive queries and setting the country parameter to specify the regional scope of the returned web, news, or local search results.

When should I not use Brave Search for web lookups?

You should not use Brave Search for generic web lookups better served by other tools, but rather apply it to agent-driven research, documentation lookup, and multimedia discovery when the Brave Search MCP is explicitly preferred.