小菲学姐文科辅导

• Outcome: Review
• Latest audit: May 29, 2026, 9:34 PM
• Version:
  3.5.0

Overview

This looks like a tutoring companion, but it includes AI-detection evasion language and broad automatic/persistent learning-profile behavior that users should review before installing.

Install only if you are comfortable with a broad learning companion that may keep student progress and emotional-profile context. Avoid using the polishing feature to hide AI-generated schoolwork, and check whether your OpenClaw setup lets you disable broad triggers, daily reminders, long-term memory, and parent-report sharing.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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

Findings (5)

Context-Inappropriate Capability

Medium

• Confidence: 96% confidence
• Finding: The skill explicitly markets '降低AIGC检测率 / AIGC Detection Reduction' as a service, which goes beyond legitimate tutoring or proofreading and facilitates evasion of AI-content detection systems. In an education context, this is especially problematic because it can be used to conceal AI-generated assignments and undermine academic integrity controls.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger keyword list is very broad and consists of common educational terms such as 'review', 'textbook', 'essay help', and 'English grammar'. In agent platforms that route or auto-invoke skills based on keyword matching, this can cause the skill to activate unintentionally, leading to prompt hijacking of unrelated study conversations and degraded user control over which skill is used.

Natural-Language Policy Violations

High

• Confidence: 98% confidence
• Finding: Advertising assistance to 'reduce AIGC detection risk' is an evasion-oriented capability expressed in natural language and directly encourages misuse. This can help users disguise AI-generated academic work or other synthetic content to bypass institutional review, making the skill materially more dangerous than ordinary writing assistance.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The trigger list contains very broad everyday terms such as "文科", "诗词", "作文", and "英语", which are likely to appear in many normal conversations unrelated to this specific skill. This can cause unintended activation or routing to the skill, creating prompt-injection exposure and confusing users by invoking the tutor in contexts where it was not explicitly requested.

Ssd 2

Medium

• Confidence: 97% confidence
• Finding: The polishing service is framed not merely as stylistic improvement but as paraphrased assistance to make AI-generated text less detectable, which is a misuse-enabling feature. In a student tutoring skill, this context increases risk because the most foreseeable use is submission fraud or circumvention of classroom authenticity checks.

VirusTotal

66/66 vendors flagged this skill as clean.