1. User asks a question requiring web research, content extraction, crawling, or site mapping.
2. The skill analyzes the user's intent and selects the appropriate Tavily API endpoint (search, extract, crawl, map, or research).
3. The skill constructs a JSON payload with the necessary parameters for the selected endpoint.
4. The skill uses the Task tool to invoke the `tavily-fetcher` sub-skill, passing the endpoint and JSON payload.
5. The `tavily-fetcher` sub-skill executes the Tavily API call.
6. The skill receives the results from the Tavily API.
7. The skill returns the results to the user.