ClawHub
Back to skill

Security audit

Bank App Wealth Advisor

Security checks across malware telemetry and agentic risk

Overview

This is a finance advisory reference skill with clear warnings and no tool access, but users should treat its sample recommendation logic as regulated financial decision-support material.

Install only if you need educational or internal reference material for bank wealth-advisory workflows. Do not use its sample product-matching code or scripts directly with customers without licensed financial-professional review, compliance approval, suitability checks, and current regulatory validation.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (3)

Intent-Code Divergence

High
Confidence
97% confidence
Finding
The skill explicitly claims it does not provide financial advice, yet it contains concrete product-recommendation logic and advisory scripts that can directly influence investment decisions. In a banking context, this mismatch can mislead deployers or users into treating regulated advisory behavior as harmless educational content, bypassing proper compliance, suitability, and human-review controls.

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The file states that no executable code is included, but it embeds working Python logic for matching products to customer profiles. Even if presented as an example, this discrepancy creates trust and governance risk: downstream systems or users may copy, run, or operationalize code that was declared absent, undermining review assumptions and control boundaries.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The manifest advertises 'advisory-only' and 'reference only' behavior, but the embedded code generates ranked product matches and recommendation reasons, which are actionable outputs. This inconsistency is dangerous because integrators may grant the skill lower scrutiny or fewer safeguards than warranted, despite it functioning as a decision-support component in a high-risk financial domain.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.