Security audit

Universal Translate Tool Free

Security checks across malware telemetry and agentic risk

Overview

This translation skill is mostly a Markdown instruction file, but it asks for command execution and file write capabilities that are not well scoped for a simple translation tool.

Review this before installing. Use it only for non-sensitive translation tasks, and do not grant command execution unless you have a specific need. For file translation, confirm the exact input and output paths and avoid translating private or credential-bearing files until the skill documents non-destructive writes and clearer data handling.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (5)

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The skill documentation says both that no extra API key is required and that the skill cannot be used without an API key. This inconsistency can mislead users and integrators about trust boundaries, dependencies, and whether external services may be involved, increasing the chance of unsafe deployment or unexpected data exposure.

Intent-Code Divergence

Medium

Confidence: 86% confidence
Finding: The skill is presented as a narrow free translation tool, but the implementation description introduces generic create/query/modify/delete operation modes. That broader capability description can normalize or justify actions beyond translation, creating ambiguity that may be abused to perform unintended state-changing operations.

Context-Inappropriate Capability

Medium

Confidence: 93% confidence
Finding: A personal translation skill should not normally require shell execution, yet the skill explicitly requests exec capability. That expands the attack surface substantially because a prompt-triggered translation workflow could be turned into command execution, file discovery, or other host-side actions unrelated to translation.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger condition is extremely broad, covering general needs like translation, language conversion, and localization. Overbroad activation criteria increase the chance that the skill is invoked in unrelated conversations, which is more dangerous here because the skill also has read/write/exec-adjacent capabilities.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The documentation describes reading a user file and saving a translated version, but it does not warn about file modification risks, overwrites, path handling, or accidental exposure of sensitive content. In a skill with write capability, missing safety guidance can lead to unintended changes to user data or processing of sensitive files without informed consent.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.