ClawHub
Back to skill

Security audit

US Tax Return Review-1040

Security checks across malware telemetry and agentic risk

Overview

This is a local tax-return review skill that writes clearly disclosed local reports containing sensitive tax data, so users need to handle the output files carefully.

Install only if you are comfortable processing tax information locally. Run it in a private directory, avoid cloud-synced or shared output folders, use --skip-docx if you do not need a Word report, restrict file permissions where possible, and delete generated reports when they are no longer needed.

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 (2)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The script persists tax-return review results to Markdown, JSON, and optionally DOCX files in a local output directory, and those artifacts include highly sensitive financial and tax information. In a tax-review skill, this is especially risky because the outputs can contain AGI, wages, credits, refund/owed amounts, and other return details that may remain on disk, be synced, or be exposed to other users or processes without the operator realizing sensitive data was written.

Missing User Warnings

Low
Confidence
84% confidence
Finding
The script reads sensitive tax-return inputs and immediately creates an output directory for artifacts without any clear privacy or retention warning. While directory creation itself is not dangerous, in this context it facilitates silent local persistence of taxpayer data and increases the chance that sensitive records are stored in insecure or unexpected locations.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.