Voice-Matched Content System

• Outcome: Review
• Latest audit: May 28, 2026 at 4:39 AM UTC
• Version:
  2.0.0

Overview

The skill’s voice-matching purpose is coherent, but it asks for potentially private writing samples and instructs profile update logging without clear user-facing privacy controls.

Install only if you are comfortable providing writing samples for style analysis. Use short excerpts, remove names, secrets, client details, and third-party messages, and avoid relying on the profile logging behavior unless you have a clear way to review and delete stored profile data.

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 (2)

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill asks users to provide multiple writing samples, including emails, texts, and casual messages, which may contain sensitive personal, professional, or third-party data. Because there is no user-facing privacy warning, minimization guidance, or redaction instruction, users may overshare private content that is then processed unnecessarily.

Ssd 3

Medium

• Confidence: 97% confidence
• Finding: The instruction to log all voice profile updates creates persistent storage of both user-provided writing characteristics and inferred behavioral traits such as confidence zones, vocabulary fingerprints, and editing instincts. Retaining this profiling data increases privacy risk, expands the blast radius of any data leak, and may conflict with data-minimization expectations if users did not clearly consent to persistence.

VirusTotal

66/66 vendors flagged this skill as clean.