websearchapi_skill

Overview

This skill provides a multi-mode web search adapter built on SearchAPI (Google) to fetch live web, news, finance, video, maps, hotels, and flights data. It returns structured JSON that an agent can parse directly and includes automatic network retry logic for improved reliability. The tool is lightweight and easy to deploy into an agent toolchain.

How this skill works

The skill calls SearchAPI endpoints with configurable parameters (type, language, region, number of results) and normalizes responses into a consistent JSON schema containing title, link, snippet, source, and metadata. It retries network requests on transient failures and supports a JSON output flag for machine consumption. Configuration commands let you set API key, defaults, and retry behavior from the CLI.

When to use it

• When an agent needs up-to-date web search results for answering queries.
• To gather recent news articles or monitor topic mentions in real time.
• To fetch finance quotes, company news, or stock-related links for investment workflows.
• To discover local places, hotels, or flight options during travel planning.
• When you need structured JSON output suitable for programmatic parsing and downstream processing.

Best practices

• Store the SearchAPI key securely and configure it via the provided CLI command before use.
• Prefer JSON output (--json) when integrating with agents or pipelines to avoid text parsing errors.
• Set appropriate --num, --lang, and --gl defaults to improve relevance for your region and use case.
• Increase retry count for unstable networks or high-latency environments to reduce transient failures.
• Use finance searches with ticker symbols (e.g., AAPL) and use English for broader financial data coverage.

Example use cases

• Agent fetches latest news about an emerging event and summarizes key developments.
• Customer support bot retrieves relevant web articles to enrich answers with citations.
• Travel assistant finds nearby hotels and flights for a user-specified destination and dates.
• Portfolio assistant queries finance results for tickers, then aggregates snippets and links for reports.
• Research agent runs web and academic-style searches and returns a structured list of sources.