监管最新动态追踪（金融监管总局、证监会、央行）

• Outcome: Pass
• Latest audit: May 29, 2026, 8:19 PM
• Version:
  1.0.0

Overview

The skill appears to monitor public regulatory news and summarize it, with some scoping and disclosure gaps but no evidence of hidden, destructive, or credential-seeking behavior.

Install if you want an agent to retrieve and summarize public Chinese financial regulatory updates. Review the configured sources and outputs, confirm before exporting files, and avoid using it for unrelated regulatory or private-document monitoring unless the scope is made explicit.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (3)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 84% confidence
• Finding: The skill instructs execution of a network-enabled crawler and mentions output support beyond simple chat responses, but it declares no permissions. This creates a transparency and governance gap: operators and users cannot accurately assess that the skill will make outbound requests and may write files, which can bypass expected review controls or policy enforcement.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 89% confidence
• Finding: The documented behavior promises title-plus-summary monitoring, but the instructions expand scope to retrieving body content and attachment links while omitting claimed Word-generation support and broader category coverage. This mismatch is dangerous because users and reviewers may approve the skill for a narrower, lower-risk purpose while it actually collects and exposes more content than expected, increasing data-handling and trust risks.

Vague Triggers

Medium

• Confidence: 81% confidence
• Finding: The trigger language is extremely broad and says the skill must be used even for generic requests about recent regulatory news. Overbroad triggering can cause inappropriate activation on ordinary queries, leading to unnecessary network access, unexpected tool use, and reduced user control over whether external retrieval occurs.

VirusTotal

66/66 vendors flagged this skill as clean.