brave-search

The brave-search skill provides a lightweight, API-driven interface for web search and content processing. Designed for AI agents that require current information without the overhead of heavy, interactive browser automation, this tool allows for efficient querying of the web. It is ideal for developers, researchers, and automated systems that need to retrieve API references, verify facts, or parse web pages into structured markdown format for further reasoning or context injection.

The skill leverages the Brave Search API to return high-quality, privacy-focused search results. Users can perform basic queries, request extended result sets, and opt-in to full-text extraction. By converting web pages into markdown, it simplifies the ingestion of documentation or news for large language models, ensuring that the retrieved data is clean and actionable. It serves as a foundational component for agents acting as research assistants, technical monitors, or automated content summarizers.

• Performs programmatic headless web searches via the Brave Search API.
• Supports adjustable result counts to balance between breadth and performance.
• Includes optional full-text extraction, converting web content directly into markdown for immediate LLM processing.
• Provides an output format that separates titles, links, snippets, and page content, facilitating easy data parsing by other agentic workflows.
• Requires a valid BRAVE_API_KEY environment variable for authentication and access to the search infrastructure.
• Designed for command-line integration and script-based automation, making it compatible with various agent development environments.
• Use for fetching technical documentation, gathering real-time data, or performing comprehensive background research on specific URLs.
• Constraints include reliance on external API availability and the quality of the Brave Search index; users should handle rate limits associated with their specific API tier.