Google Cloud Translate

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:41 PM UTC
• Version:
  1.0.2

Overview

This is a disclosed Google Cloud Translate skill that uses Membrane for authentication and API access, with some broad but visible proxy-request capability users should control carefully.

Install only if you trust Membrane and the npm CLI package, review the Google permissions granted during connection, and require explicit approval before using raw proxy requests that create, update, delete, or send sensitive content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The skill is framed as a Google Cloud Translate integration, but it documents a much broader capability: creating connections and issuing arbitrary proxied HTTP requests, including state-changing methods. That expands the operational scope beyond translation and can enable unintended access or modification of external resources if the agent follows the documentation too literally.

Context-Inappropriate Capability

Medium

• Confidence: 93% confidence
• Finding: Including a generic raw HTTP proxy in a narrowly branded integration gives the agent a powerful escape hatch that bypasses the safety of curated actions. Even if intended for flexibility, this increases the chance of data exfiltration, overbroad API access, or misuse against endpoints unrelated to the user's requested translation task.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The proxy-request guidance tells the agent how to send direct requests to an external API but does not warn that user-provided content may be transmitted off-platform. In a translation context, text commonly contains sensitive business or personal data, so omission of a disclosure or consent checkpoint raises data-handling risk.

VirusTotal

65/65 vendors flagged this skill as clean.