ClawHub
Back to skill

Security audit

Refund Radar

Security checks across malware telemetry and agentic risk

Overview

Refund Radar is a coherent local bank-statement audit skill, but its generated reports and saved state can contain sensitive financial history.

Install only if you are comfortable giving the tool bank or card transaction history. Keep the generated ~/.refund_radar reports private, delete them when no longer needed, avoid entering full card or account numbers in templates, and verify any separately installed refund_radar Python code before running it on real statements.

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 (2)

Vague Triggers

Medium
Confidence
87% confidence
Finding
The trigger phrase "analyze my credit card transactions" is broad natural-language wording that could match ordinary user requests outside a tightly scoped invocation pattern. In an agent environment, this can cause the skill to activate unexpectedly on sensitive financial conversations and process bank or card data when the user did not clearly intend to invoke this specific skill.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger phrase "find recurring charges in my statement" is generic and likely to overlap with normal conversational requests for financial help. Because this skill handles highly sensitive banking data and writes reports to disk, accidental invocation increases privacy risk and may expose or persist transaction information without sufficiently explicit user intent.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.