AI 企业画像

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:29 PM UTC
• Version:
  0.1.3

Overview

This appears to be a legitimate AI company research skill that calls a disclosed external API, with some overbroad wording and missing privacy guidance but no hidden, destructive, or persistent behavior.

Install this only if you are comfortable sending AI-company research queries to the Jiqizhixin API using your JQZX_API_TOKEN. Treat the token as secret, avoid using the skill for confidential acquisition, investment, or competitive targets unless that disclosure is acceptable, and consider narrowing the trigger language locally so it is used for enterprise research rather than every general AI question.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• 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 (2)

Natural-Language Policy Violations

Medium

• Confidence: 81% confidence
• Finding: The description says AI-industry questions 'must' use this skill and forbids relying on model memory for concrete facts, which attempts to override normal agent/user choice and can steer queries to an external provider regardless of user intent. This creates a policy-style coercion that may cause unnecessary external data disclosure and reduces the agent's ability to apply safer or more appropriate alternatives.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill instructs users to configure `JQZX_API_TOKEN` and use direct API/script access, but it does not warn that the token is sensitive, that requests go to an external service, or that query contents may be transmitted off-platform. In enterprise research contexts, searched company names, competitive targets, or strategic topics can themselves be sensitive business information.

VirusTotal

66/66 vendors flagged this skill as clean.