Freepik

Security checks across malware telemetry and agentic risk

Overview

This skill appears to be a straightforward Freepik API integration, with expected API-key use, Freepik network calls, and local saving of generated media.

Install this only if you are comfortable giving the agent access to a Freepik API key. Prompts, selected media URLs or files, and related request metadata may be sent to Freepik and may consume account credits. Review saved outputs under ~/.freepik/sessions/ if you do not want generated media retained locally.

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 (2)

Intent-Code Divergence

Medium

Confidence: 94% confidence
Finding: The skill’s security guidance explicitly tells the agent to present result URLs directly, but later sections instruct it to use curl to download files into local session storage. That inconsistency can cause the agent to retain generated or downloaded media unnecessarily, increasing privacy and data-retention risk and undermining the stated safety boundary.

Missing User Warnings

Medium

Confidence: 79% confidence
Finding: The setup instructions tell users to configure an API key but do not clearly warn that prompts and uploaded media will be transmitted to Freepik's external service. In an agent skill context, users may submit sensitive text or images through natural-language requests, so lack of disclosure can lead to unintended third-party data exposure and privacy/compliance issues.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal