Perspective Control System

• Outcome: Pass
• Latest audit: Jun 10, 2026 at 4:13 PM UTC
• Version:
  1.0.1

Overview

This is a text-only narrative writing guidance skill with no executable code or hidden data access, though it may strongly steer POV style and Vietnamese-flavored examples.

Install this if you want strict POV discipline applied to narrative scenes. Be aware it may over-apply to general fiction writing and may nudge outputs toward Vietnamese example phrasing unless you explicitly ask the agent to keep your requested language and style.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The description says to use the skill when writing or reviewing "any narrative scene," which is an unusually broad activation condition. That can cause the skill to trigger outside narrowly intended contexts, overriding user preferences or more relevant instructions and creating prompt-scope confusion rather than direct code-execution risk.

Natural-Language Policy Violations

Medium

• Confidence: 88% confidence
• Finding: The skill's examples and rule wording are heavily Vietnamese-centric, which can steer outputs into Vietnamese or imply a language default without user consent. This is mainly a policy/usability vulnerability: it can cause unauthorized language switching, reduced instruction fidelity, and confusion when the user requested another language.

VirusTotal

58/58 vendors flagged this skill as clean.