ClawHub

Tavily Search

Security checks across malware telemetry and agentic risk

Overview

This Tavily skill is a straightforward external web-search integration, but users should understand that queries, URLs, crawls, and connection actions go through Maton and Tavily.

Install only if you trust Maton and Tavily with the searches, URLs, crawl targets, extracted content, and API access you provide. Avoid submitting secrets, confidential internal URLs, regulated data, or sensitive research targets unless external sharing is intended, and list/check connections before deleting or selecting one.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill clearly enables external web search, extraction, crawling, and research through Maton/Tavily, but the description does not explicitly warn that user prompts, URLs, and fetched content are transmitted to third-party services. In an agent context, this can cause inadvertent disclosure of sensitive user data because operators may rely on the skill description to understand privacy boundaries.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The connection-management section documents creating, listing, retrieving, and deleting remote API-key connections, but it lacks a clear warning that these actions manipulate stored third-party credentials and that deletion may be destructive. In practice, this can lead to accidental credential exposure, misuse of the wrong connection, or unintended service disruption if a connection is deleted without confirmation.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal