Tavily Web

Tavily Web Skill

Trigger Conditions & Endpoint Selection

Choose Tavily endpoint based on user intent:

• search: Need to "search web / latest info / find sources / find links"
• extract: Given URL(s), need to extract/summarize content
• crawl: Need to traverse site following instructions and scrape page content
• map: Need to discover site page list/structure (without full content or metadata only)
• research: Need structured research output following given output_schema

Recommended Architecture (Main Skill + Sub-skill)

This skill uses a two-phase architecture:

1. Main skill (current context): Understand user question → Choose endpoint → Assemble JSON payload
2. Sub-skill (fork context): Only responsible for HTTP call execution, avoiding conversation history token waste

Execution Method

Use Task tool to invoke tavily-fetcher sub-skill, passing command and JSON (stdin):

Task parameters:
- subagent_type: Bash
- description: "Call Tavily API"
- prompt: cat 
  { ...payload... }
  JSON

Payload Examples (Based on Provided curl)

1) Search the web

cat ",
  "chunks_per_source": 3,
  "extract_depth": "basic",
  "include_images": false,
  "include_favicon": false,
  "format": "markdown",
  "timeout": "None",
  "include_usage": false
}
JSON

3) Crawl webpages

cat  `.env`):

1. Environment variable: `TAVILY_API_KEY`
2. `.env` file: Place in `.env`, can copy from `.env.example`

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [BenedictKing](https://github.com/BenedictKing)
- **Source:** [BenedictKing/benedictking-skills](https://github.com/BenedictKing/benedictking-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.