ClawHub
Back to skill

Security audit

Endowment Effect

Security checks across malware telemetry and agentic risk

Overview

This is a text-only behavioral economics coaching skill with no code execution or data access, though users should apply its free-trial and personalization advice ethically.

Install only if you want behavioral-economics guidance for negotiation, pricing, or product design. When using its SaaS trial or personalization suggestions, require clear user consent, easy data deletion/export, salient renewal terms, and cancellation that is no harder than signup.

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
87% confidence
Finding
The skill explicitly recommends using free trials, personalization, and data import to increase user attachment and willingness-to-pay, which can encourage manipulative product design without sufficient privacy, consent, or data-minimization guardrails. Although it later mentions an ethical check and FTC click-to-cancel risk, the operational guidance still normalizes behavioral manipulation and collection/use of user data as a retention lever.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.