ClawHub
Back to skill

Security audit

Ask a Tax Nerd

Security checks across malware telemetry and agentic risk

Overview

This tax helper is not malicious, but it needs Review because its tax-advice scope is broad and may understate jurisdiction and professional-advice limits.

Review before installing. Use it only for general tax organization or preliminary research, confirm your jurisdiction before relying on any answer, and have a CPA, enrolled agent, or tax attorney review material filing positions or complex deductions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (3)

Vague Triggers

High
Confidence
92% confidence
Finding
The skill’s trigger scope is excessively broad and explicitly says to activate for nearly any casual tax mention, even when the user may not be asking for tax analysis. That creates a routing/prompt-quality vulnerability: the agent can be pushed into authoritative tax guidance unnecessarily, increasing the chance of irrelevant, overconfident, or unsafe advice in sensitive financial contexts.

Vague Triggers

Medium
Confidence
90% confidence
Finding
Defaulting to US federal tax answers when jurisdiction is unclear is dangerous because tax rules are highly jurisdiction-specific, and the skill itself references non-US authorities like HMRC. A user with a non-US question could receive incorrect but confident US guidance, which is especially risky in tax matters where filing, reporting, and deadlines vary materially by country.

Natural-Language Policy Violations

Medium
Confidence
88% confidence
Finding
The guide explicitly discourages a user-facing statement that clarifies the model is not a tax professional and is not providing professional advice. In a tax skill, this can materially increase user overreliance on AI output, especially when the same document encourages confident answers and minimizing caution, which raises the chance that users act on incomplete or jurisdiction-specific guidance without realizing its limits.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal