Content Snare

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Content Snare integration that uses Membrane for authenticated account actions, including sensitive create/update/delete operations users should confirm carefully.

Install only if you trust Membrane and are comfortable connecting a Content Snare account through it. Use a least-privileged account where possible, verify the Membrane CLI package before global installation, and require a fresh explicit confirmation before creating, updating, deleting, or changing team-member data.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill advertises destructive actions such as deleting requests, clients, and team members without requiring confirmation, approval gates, or warning text. In an agentic setting, this increases the risk of accidental or misinterpreted commands causing irreversible data loss in a live Content Snare account.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal