skylar-serper_skill

Overview

This skill lets you perform Google searches programmatically via the Serper.dev API, returning clean, structured results without scraping. It supports result limiting, geographic (gl) and language (hl) targeting, and is designed for integration into agents or automation workflows. An API key stored in SERPER_API_KEY is required to run searches.

How this skill works

You call the search tool with a query and optional parameters (num, gl, hl). The tool sends the request to Serper.dev and returns parsed search results, including links and summaries, in a consistent format. It handles rate-limited result counts (1–100) and surfaces the API response for downstream processing.

When to use it

• Automating web research or fact-finding for agents and bots
• Collecting search results filtered by country or language
• Fetching a fixed number of top search results for summarization
• Replacing ad-hoc scraping with an API-backed search source
• Integrating web search into data pipelines or assistants

Best practices

• Set SERPER_API_KEY in your environment to authorize requests
• Limit num to what you need (default 5) to reduce latency and cost
• Use gl and hl to target local results and improve relevance
• Validate and normalize results before using them in decision logic
• Handle API errors and rate limits gracefully in your agent

Example use cases

• Agent that summarizes the top 5 search results for a user query
• Research workflow that collects country-specific news or pages
• Content monitoring that pulls search results for keyword tracking
• Enrichment step that fetches web citations before answering user questions
• Batch job that exports structured search results for analysis