Use Case Skills

Outcome-oriented Tavily skills for common research workflows.

This repo is the use-case workflow counterpart to the Tavily endpoint skills in tavily-ai/skills. Endpoint-specific execution details should come from those companion skills; this pack focuses on how to plan, navigate, verify, and format high-value Tavily research workflows.

Installation

Install the endpoint skills from tavily-ai/skills alongside this repo. Use those skills for setup, authentication, and tool-specific execution details.

This repo intentionally organizes skills by broad use case instead of by endpoint. Each skill chooses the right capability pattern internally:

• search-oriented discovery followed by targeted extraction for fast source-grounded research
• site mapping before extraction when a known domain needs URL discovery
• scoped crawling only when many pages from a known section are needed
• research only for deep report-style synthesis

Each skill applies Tavily best practices: short search queries under 400 characters, source filtering before extraction, map-before-crawl planning, conservative collection scope, clear research goals, citation hygiene, and explicit coverage gaps.

Default execution discipline:

• Start with a small focused set of search queries, not the full possible query list.
• Keep each search query under 400 characters.
• Extract only the strongest sources needed for the deliverable.
• Expand with additional searches only for named evidence gaps.
• Use map only after search shows a known site has useful but buried pages.
• Use crawl only after map identifies a narrow section worth collecting.
• Stop when the output template can be filled with cited evidence and clear gaps.

Skills

Skill	Primary use case	Default capability pattern
vendor-risk-kyc-screening	Vendor onboarding, KYC, adverse media, sanctions, and supplier risk	discovery, source verification, targeted extraction
sales-account-intelligence	Sales prep, account briefs, buyer research, and trigger events	account discovery, source verification, targeted extraction
product-competitor-intelligence	Product/SKU discovery, pricing/spec extraction, and competitor intelligence	discovery, site navigation, scoped collection
threat-intelligence-enrichment	CVE, IOC, advisory, exploit, mitigation, and incident enrichment	exact-identifier discovery, authoritative extraction
investment-research-briefs	Concise investor briefs, sector snapshots, and risk/catalyst memos	research synthesis, source verification
academic-scientific-research	Papers, scientific literature, methods, evidence summaries	scholarly discovery, evidence extraction
migrate-to-tavily	Migrate existing Exa/Firecrawl/Perplexity/Parallel integrations to Tavily	code mapping, in-place rewrite
fill-missing-fields	Find and fill in missing fields on an existing list of companies, people, or products	structured per-row research
build-entity-list	Build a deduplicated list of entities matching given criteria	structured list research
watch-for-changes	Watch a page, site, or topic for changes on a recurring schedule	extract/search + scheduled orchestration
past-research-search	Find a past research run by topic instead of by run ID	local run index

Design

These skills are designed to compose endpoint-oriented skills from tavily-ai/skills into marketable workflows that map to common customer needs.