ClawHub
Back to skill

Security audit

Market Sentiment

Security checks across malware telemetry and agentic risk

Overview

This skill is a local market-sentiment CSV analyzer that gives finance-related suggestions, but it does not access accounts, trade, exfiltrate data, or persist in the background.

Install only if you want a local market-sentiment decision-support tool. Review the stock-planner mapping before automating workflows, and do not let the position or threshold suggestions drive real trades without independent review and suitability checks.

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 (4)

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The README shows a contradictory integration example: a strong/overheated market output is mapped to `stock-planner: --market bear`. If users automate around this example, they may feed the opposite market regime into downstream planning logic and trigger systematically wrong trading behavior. In this skill context, the issue is more dangerous because the tool is explicitly intended to drive another agent/tool via command substitution, so documentation errors can become operational misconfigurations rather than harmless text mistakes.

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
This script does more than descriptive sentiment analysis: it generates concrete portfolio-position and threshold recommendations such as how many holdings to keep and when to add or reduce exposure. In an agent skill context, that can directly influence financial decisions and downstream automation, especially because the output is formatted for integration with other tools.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The disclaimer says the tool is not investment advice, but the code later emits direct action-oriented recommendations about position sizing, adding/reducing holdings, and stop-loss thresholds. This mismatch can mislead users or integrators into overtrusting the output while the author attempts to disclaim responsibility, increasing the risk of unsafe financial reliance.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The trigger list includes broad terms such as '市场情绪', '大盘环境', 'sentiment', and 'market state' that can match ordinary financial discussion, causing the skill to activate outside clearly intended contexts. Over-broad invocation can route unrelated user prompts into this skill, leading to unintended data handling, incorrect automation in a trading workflow, or prompt-surface expansion for downstream tools.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.