Tavily Best Practices

Security checks across static analysis, malware telemetry, and agentic risk

Overview

The skill is largely documentation for a web-search/crawl SDK (Tavily) and appears functionally consistent, but the runtime instructions require placing an API key into a user config path (~/.claude/settings.json) and reference environment variables not declared in the registry metadata—an incoherence you should review before installing or using.

What to consider before installing or using this skill: - The documentation requires a Tavily API key (TAVILY_API_KEY) and instructs adding it to ~/.claude/settings.json, but the skill metadata declares no required environment variables or config paths — confirm you are comfortable storing that key in your Claude settings file and that this is the intended place for it. - The skill is instruction-only (no code to execute from the registry), which lowers install risk, but it points you to install third‑party packages (pip/npm). Only install those packages from official package registries and verify package names (tavily-python, @tavily/core) come from the official Tavily organization. - The functionality (search, extract, crawl) legitimately needs an API key and network access; these operations will fetch web pages and could retrieve sensitive or private content if you target internal URLs — apply domain filters, limits, and respect robots.txt as the docs advise. - Ask the publisher or registry maintainer to correct the metadata: required env vars (TAVILY_API_KEY) and any config paths should be declared. If the skill came from an unknown source (homepage/source unknown here), verify the origin before providing credentials. - If you want stronger safety: test in a sandbox account/project with a scoped Tavily key, monitor API usage, and avoid using your primary production credentials until you confirm behavior.

Static analysis

Static analysis findings are pending for this release.

VirusTotal

No VirusTotal findings

View on VirusTotal

Risk analysis

No visible risk-analysis findings were reported for this release.