太子三省六部团队

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:37 PM UTC
• Version:
  0.1.5

Overview

This is a markdown-only business-advisory persona skill with disclosed courtly style and no code, credential access, persistence, or hidden execution behavior.

Install this if you want a formal, court-themed business advisory style. Avoid pasting confidential company, client, contract, financial, personal, or credential data unless you have permission and a secure workflow, and independently verify high-stakes legal, financial, hiring, or business decisions.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Natural-Language Policy Violations

Medium

• Confidence: 93% confidence
• Finding: The skill hard-codes a deferential form of address ('殿下' / '主上') and discourages direct second-person address unless the user opts out. This is a real prompt-quality and autonomy issue because it imposes a social framing that may pressure or manipulate user interactions, especially in advisory contexts where neutrality and user preference should come first.

Missing User Warnings

Low

• Confidence: 89% confidence
• Finding: The usage examples invite users to submit business situations, pricing, cooperation decisions, and reviews without any warning against including confidential, personal, or regulated information. In a skill explicitly positioned for founders and operators handling complex business matters, this increases the chance that users will paste sensitive company, client, financial, or contractual data into prompts.

Natural-Language Policy Violations

Medium

• Confidence: 91% confidence
• Finding: The skill hard-codes honorific forms of address (殿下/主上) and explicitly forbids a common second-person form unless requested. This can override user preference, create unwanted social pressure or manipulation dynamics, and reduce user autonomy in the interaction, especially in a system framed as a 'private court' with strong authority cues.

VirusTotal

66/66 vendors flagged this skill as clean.