tavily-research: a Claude Code skill by Tavily AI
Official
For people who need a sourced write-up on a market, policy or technology question. Skip it if you want one quick fact or don't have Tavily's tvly CLI set up.
- Score
- 76of 100
- Rank
- #109of 239
- Installs
- 24Kon skills.sh
- Security
- Mixed 4 scanners
- Updated
- 2026
- Context
- ~132tokens always
Install
npx skills add https://github.com/tavily-ai/skills --skill tavily-research
Run it in your terminal. View the source
Using the Claude app rather than Claude Code? You can add a skill there too. How to add it in the app
What does tavily-research do?
tavily-research is a skill that builds a researched report from many sources with Tavily's tvly CLI. It asks a question only when the answer would change the work, then splits the topic into three or more subtopics that don't overlap, carrying your dates, places and limits into each. It either starts one subagent per subtopic, each writing notes under research_notes/, or runs the searches itself in parallel. Thin sections get another round of searching before writing. The report follows a bundled writer guide and is saved to reports/ when you ask for a file. Single facts go to tavily-search instead.
When should you use tavily-research?
Use it when
- You need a market or trend analysis from many sources
- You want options compared across several criteria
- You need a short literature review with links
Skip it when
- You only need one fact answered
- You don't have Tavily's CLI set up
What it needs and what it can touch
You need
- Tavily's tvly CLI, set up as the tavily-search skill describes
- The tavily-search skill
What it can do
- Runs shell commands (tvly)
- Sends search queries to an outside service (Tavily)
- Starts subagents and writes notes and reports to your working folder
From reading its instructions, not from running it.
Why it scores 76
Strongest on upkeep (100 of 100), weakest on adoption (40).
Is tavily-research safe to install?
At least one scanner raised a medium warning. Read the results below. Scanners often miss what other scanners catch, and a skill can change after a scan. Read its source before you install it.
| Scanner | Result | Date |
|---|---|---|
| Gen Agent Trust Hub | Passed | Oct 30, 2026 |
| Socket | Passed | Oct 30, 2026 |
| Snyk | Medium risk | Oct 30, 2026 |
| ZeroLeaks | Low risk | Apr 16, 2026 |
Details
- Installs
- 23,959
- Repository stars
- 490
- Last repository change
- Tokens always loaded
- about 132
- Tokens when used
- about 512
- License
- MIT
- Data confidence
- High
Did it do the job for you?
Alternatives to tavily-research
All skills for this job-
Get Academic Research SkillsAcademic Research Skills
For researchers who want a staged, checked path from research question to finished paper.
-
Get Scientific Agent SkillsScientific Agent Skills
For researchers who want Claude to run analyses with scientific Python tools in biology, chemistry or statistics.
-
Get gmgn-portfoliogmgn-portfolio
For crypto traders who want wallet holdings, profit and loss, win rate and dev launch history from GMGN.
-
Get ARIS (Auto Research In Sleep)ARIS (Auto Research In Sleep)
For ML researchers who want agents to plan experiments, review drafts with a second model and outline papers.