ClawHub
Back to skill

Security audit

trading-news-guard

Security checks across malware telemetry and agentic risk

Overview

This skill is only documentation, but its local/no-network claims conflict with API-based trading-guard instructions in a financially sensitive workflow.

Review this carefully before installing or relying on it for trading. It does not include working code, a defined data source, or a verifiable command, and its documentation disagrees about whether it uses an API or stays local. Do not wire it into automated trading without replacing the ambiguous parts with a tested, explicit implementation and clear fail-closed behavior.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (2)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The manifest and metadata repeatedly claim the skill is '100% local' with no network calls, but the body says it 'fetches' status from an API and instructs code to call `check_news()`. This mismatch can mislead users, reviewers, or policy systems into granting the skill more trust or permissions than its real behavior warrants, which is a security-relevant documentation integrity issue.

Vague Triggers

Medium
Confidence
77% confidence
Finding
The command/activation section is incomplete and ambiguous: it shows only a comment and no actual invocation or interface contract. In an automated trading context, unclear integration instructions increase the chance that the guard is never called, called incorrectly, or bypassed entirely, undermining a safety control the surrounding documentation presents as critical.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.