Snowflake

• Outcome: Review
• Latest audit: May 29, 2026 at 9:37 PM UTC
• Version:
  1.0.2

Overview

This Snowflake skill is not deceptive, but it gives an agent broad authenticated access that could change warehouse data without clear safety guardrails.

Install only if you trust Membrane and intend an agent to access Snowflake. Use a least-privilege or read-only Snowflake role where possible, prefer prebuilt Membrane actions, and require explicit approval before any write, delete, schema, permission, or administrative operation.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The skill explicitly documents raw proxy access and supports state-changing HTTP methods like POST, PUT, PATCH, and DELETE without requiring confirmation, read-only defaults, or warnings about destructive effects. In a Snowflake context, this can enable accidental or overly broad data modification, administrative changes, or destructive operations if an agent uses the proxy path directly instead of constrained actions.

VirusTotal

65/65 vendors flagged this skill as clean.