Funnel Helper

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only marketing funnel analysis skill with no executable code, persistence, credential use, or hidden authority.

Safe to install as a marketing analysis helper. Before using it, provide aggregated funnel metrics where possible and remove personal identifiers or unnecessary raw session, form, and checkout log details.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

Findings (1)

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger keywords are generic marketing terms such as "conversion," "campaign," and "optimize," which can cause the skill to activate for broad ad/marketing discussions that do not specifically require funnel diagnosis. Overbroad activation can route users into irrelevant automation or analysis paths, increasing the chance of incorrect guidance, context leakage across skills, or unintended tool usage in larger agent systems.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal