Userpilot

Security checks across malware telemetry and agentic risk

Overview

This appears to be a real Userpilot integration, but it gives an agent broad authenticated API access that could change or delete business data without built-in safeguards.

Install only if you trust Membrane and intend to let an agent operate on your Userpilot workspace. Use the least-privileged Userpilot/Membrane connection available, review the exact endpoint, method, and payload before any create/update/delete request, and revoke the Membrane connection when you no longer need it.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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)

Vague Triggers

Medium

Confidence: 87% confidence
Finding: The skill description is broad enough that an orchestrator could invoke it for generic requests about managing data or workflows, even when the user did not specifically ask to interact with Userpilot. Because this skill supports authenticated external actions and proxy requests, overbroad routing increases the chance of unintended access, queries, or state-changing operations in the wrong system.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The documentation explicitly enables raw proxied API calls with mutating HTTP methods like POST, PUT, PATCH, and DELETE, but provides no warning about confirmation requirements, scope validation, or data-loss risk. In an agent setting, this makes it easier for a model to perform irreversible changes against a live Userpilot tenant without sufficient user awareness or approval.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal