Security audit

Multi Search Engine

Security checks across malware telemetry and agentic risk

Overview

This is a search aggregation skill whose core behavior is mostly purpose-aligned, but its privacy notice incorrectly says there is no external data transmission while it sends queries to many third-party search engines.

Install only if you are comfortable with your search terms being sent to the listed third-party search engines, potentially across regions. Do not use it for secrets, credentials, internal project names, regulated data, or sensitive personal queries unless you explicitly choose the provider and understand its logging and privacy practices.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Intent-Code Divergence

Medium

Confidence: 96% confidence
Finding: The documentation states that 'all operations run locally, no external data transmission' even though the skill explicitly sends user queries and HTTP requests to multiple third-party search engines. This is a misleading privacy claim that can cause users or integrators to disclose sensitive queries under false assumptions about data handling.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The examples explicitly instruct use of web_fetch against third-party search engines, which sends user queries and possibly related metadata to external providers. In a search skill this behavior is expected, but omitting any warning or consent note can cause unintended disclosure of sensitive queries, especially when using multiple providers across different jurisdictions.

Natural-Language Policy Violations

Medium

Confidence: 79% confidence
Finding: The skill automatically routes queries to Chinese or international engines based on detected language without explicit user choice or consent. This can expose user queries to jurisdictions, providers, or censorship regimes they did not intend, creating privacy and policy risks especially for sensitive searches.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The document contains many example `web_fetch` calls to third-party search services but does not warn that user queries, metadata, and possibly sensitive research terms will be transmitted off-platform. In a search-integration skill, this is contextually expected, but the absence of disclosure can still cause unintended privacy leakage, especially when users treat examples as safe defaults.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.