My Liang Tavily Search

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

Overview

This appears to be a straightforward Tavily web-search tool, but it uses your Tavily API key, sends searches to Tavily, and has a minor package-metadata mismatch worth verifying.

This skill looks safe for its stated purpose. Before installing, confirm the publisher/version despite the metadata mismatch, provide only a Tavily API key you are comfortable using, and avoid searching for confidential information unless you are comfortable sending those queries to Tavily.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

The skill can use the Tavily account quota and permissions associated with the provided API key.

Why it was flagged

The script reads the Tavily API key from the environment and uses it as a bearer token for the Tavily search request. This is expected for the stated integration, but it is credential use.

Skill content
const apiKey = (process.env.TAVILY_API_KEY ?? "").trim(); ... "Authorization": `Bearer ${apiKey}`,
Recommendation

Use a dedicated Tavily API key if possible, store it securely as a secret or environment variable, and rotate it if you no longer trust the installation.

What this means

Search terms and selected options may be processed by Tavily outside the local environment.

Why it was flagged

The user's query and search options are sent to Tavily's external API. This is disclosed and purpose-aligned, but it is still an external data flow.

Skill content
await fetch("https://api.tavily.com/search", { method: "POST", ... body: JSON.stringify(body) });
Recommendation

Do not submit secrets, private documents, or sensitive internal queries unless Tavily's handling of that data is acceptable to you.

What this means

A model using these results could be influenced by inaccurate or adversarial web content if it treats the content as instructions rather than evidence.

Why it was flagged

The skill prints answers and web-result snippets returned by Tavily into the agent's context. Web content is untrusted and may contain misleading or instruction-like text.

Skill content
console.log(data.answer); ... const content = String(r?.content ?? "").trim(); ... console.log(`  ${content.slice(0, 300)}${content.length > 300 ? "..." : ""}`);
Recommendation

Treat search results as untrusted reference material, prefer cited sources, and do not follow instructions found inside returned webpages unless separately verified.

What this means

A metadata mismatch can make it harder to confirm that the package came from the expected publisher/version.

Why it was flagged

The embedded _meta.json owner/version differ from the supplied registry metadata, which lists a different owner ID and version 1.0.0. This is a provenance/packaging inconsistency, not evidence of malicious runtime behavior.

Skill content
"ownerId": "kn724ben7jjrcy3pebw2jbak3s8201cp", "slug": "my-liang-tavily-search", "version": "1.0.1"
Recommendation

Verify the publisher and version in the registry before installing, especially because the source is listed as unknown.