← Back to Skills Marketplace
evanydl

Tavily Research

by evanYDL · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
525
Downloads
0
Stars
5
Active Installs
1
Versions
Install in OpenClaw
/install tavily-research
Description
Comprehensive research grounded in web data with explicit citations. Use when you need multi-source synthesis—comparisons, current events, market analysis, d...
README (SKILL.md)

Research Skill

Conduct comprehensive research on any topic with automatic source gathering, analysis, and response generation with citations.

Authentication

The script uses OAuth via the Tavily MCP server. No manual setup required - on first run, it will:

  1. Check for existing tokens in ~/.mcp-auth/
  2. If none found, automatically open your browser for OAuth authentication

Note: You must have an existing Tavily account. The OAuth flow only supports login — account creation is not available through this flow. Sign up at tavily.com first if you don't have an account.

Alternative: API Key

If you prefer using an API key, get one at https://tavily.com and add to ~/.claude/settings.json:

{
  "env": {
    "TAVILY_API_KEY": "tvly-your-api-key-here"
  }
}

Quick Start

Tip: Research can take 30-120 seconds. Press Ctrl+B to run in the background.

Using the Script

./scripts/research.sh '\x3Cjson>' [output_file]

Examples:

# Basic research
./scripts/research.sh '{"input": "quantum computing trends"}'

# With pro model for comprehensive analysis
./scripts/research.sh '{"input": "AI agents comparison", "model": "pro"}'

# Save to file
./scripts/research.sh '{"input": "market analysis for EVs", "model": "pro"}' ./ev-report.md

# Quick targeted research
./scripts/research.sh '{"input": "climate change impacts", "model": "mini"}'

Parameters

Field Type Default Description
input string Required Research topic or question
model string "mini" Model: mini, pro, auto

Model Selection

Rule of thumb: "what does X do?" -> mini. "X vs Y vs Z" or "best way to..." -> pro.

Model Use Case Speed
mini Single topic, targeted research ~30s
pro Comprehensive multi-angle analysis ~60-120s
auto API chooses based on complexity Varies

Examples

Quick Overview

./scripts/research.sh '{"input": "What is retrieval augmented generation?", "model": "mini"}'

Technical Comparison

./scripts/research.sh '{"input": "LangGraph vs CrewAI for multi-agent systems", "model": "pro"}'

Market Research

./scripts/research.sh '{"input": "Fintech startup landscape 2025", "model": "pro"}' fintech-report.md
Usage Guidance
What to consider before installing: - This skill will send your research queries to Tavily (https://mcp.tavily.com). Only use it if you trust Tavily to handle the content you submit. - The script tries to reuse OAuth tokens from ~/.mcp-auth and will also open a browser OAuth flow (via npx mcp-remote) if no token is found. Expect a transient npm download when the OAuth helper runs. - The script expects command-line tools (jq, curl, base64, date, find, npx) but the skill metadata doesn't list these requirements — ensure they exist before running. - The SKILL.md suggests setting TAVILY_API_KEY in ~/.claude/settings.json as an option; the script only honors the TAVILY_API_KEY environment variable or tokens in ~/.mcp-auth. If you prefer to avoid the OAuth flow, set TAVILY_API_KEY in your environment beforehand. - Because source/homepage are not provided in the registry metadata, exercise normal caution: review the script (you already did) and confirm the mcp domain and issuer string (script checks iss == "https://mcp.tavily.com/") match the official service you intend to use. - Avoid sending highly sensitive secrets or proprietary data through the skill unless you have verified and accepted Tavily's data handling policies. If you want higher assurance, ask the publisher for a homepage or signed source, or run the script in an isolated environment first.
Capability Analysis
Type: OpenClaw Skill Name: tavily-research Version: 1.0.0 The skill provides a research interface using the Tavily API and MCP (Model Context Protocol). The script `scripts/research.sh` implements a legitimate OAuth flow, including searching for existing Tavily-issued tokens in `~/.mcp-auth/` and validating them by decoding the JWT payload to check the issuer (`https://mcp.tavily.com/`). It uses `npx` to run the standard `mcp-remote` tool for authentication and communicates only with official Tavily endpoints. The logic is transparent, well-documented in `SKILL.md`, and lacks any indicators of data exfiltration or malicious intent.
Capability Assessment
Purpose & Capability
The skill claims to perform web-grounded research and uses Tavily's MCP endpoint (https://mcp.tavily.com/mcp) to do so. Requiring a Tavily token or API key and consulting an MCP auth cache (~/.mcp-auth) is coherent with that purpose; no unrelated services or credentials are requested.
Instruction Scope
Runtime instructions and the included script stick to conducting research via Tavily's MCP. The script reads ~/.mcp-auth for *_tokens.json and may launch an OAuth flow (via npx mcp-remote). Minor issues: SKILL.md suggests adding TAVILY_API_KEY to ~/.claude/settings.json as an alternative, but the script does not parse that file (it respects the TAVILY_API_KEY environment variable). The script also assumes utilities (jq, curl, base64, date, find, npx) are available but the skill manifest does not declare required binaries.
Install Mechanism
There is no install spec (instruction-only with an included helper script), so nothing is permanently installed by the skill bundle. The script may invoke npx at runtime (which transiently fetches an npm package) but there is no packaged download that writes arbitrary archives to disk.
Credentials
The script legitimately needs a Tavily credential (TAVILY_API_KEY or an OAuth token in ~/.mcp-auth). However, the skill metadata did not declare a primary credential or required env var; the SKILL.md documents the API key option but the registry fields omit it. The script's access is limited to the user's home auth cache (~/.mcp-auth) and the Tavily endpoint, which is proportional to the purpose but should be declared explicitly.
Persistence & Privilege
The skill is not force-included (always: false), does not attempt to modify other skills or global agent configuration, and does not request elevated or persistent system privileges. OAuth flow spawns a short-lived npx process but that's normal for an OAuth helper.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install tavily-research
  3. After installation, invoke the skill by name or use /tavily-research
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
tavily-research v1.0.0 – Initial public release. - Comprehensive research with web-cited synthesis and multiple models (mini, pro, auto). - Supports OAuth authentication (browser-based) or API key via settings. - Flexible CLI usage: input topic as JSON; optional output to file. - Model choice for either quick overviews or deep analysis. - Detailed setup and usage instructions provided in SKILL.md.
Metadata
Slug tavily-research
Version 1.0.0
License MIT-0
All-time Installs 8
Active Installs 5
Total Versions 1
Frequently Asked Questions

What is Tavily Research?

Comprehensive research grounded in web data with explicit citations. Use when you need multi-source synthesis—comparisons, current events, market analysis, d... It is an AI Agent Skill for Claude Code / OpenClaw, with 525 downloads so far.

How do I install Tavily Research?

Run "/install tavily-research" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Tavily Research free?

Yes, Tavily Research is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Tavily Research support?

Tavily Research is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Tavily Research?

It is built and maintained by evanYDL (@evanydl); the current version is v1.0.0.

💬 Comments