Dlazy Audio Generate

• Outcome: Review
• Latest audit: Jun 2, 2026 at 10:26 AM UTC
• Version:
  1.2.0

Overview

This appears to be a real dLazy audio-generation integration, but its instructions are broader than an audio-only skill and deserve review before use.

Review before installing. Use it only if you trust dLazy with the prompts and media you provide, confirm any local file before upload, prefer explicit audio-specific commands, and consider using npx or a pinned reviewed CLI version instead of a broad global install.

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 (3)

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The skill is marketed as audio generation, but its documented behavior includes generic piping plus image/video examples and instructions that can route local files and upstream outputs into other media workflows. This broadens the effective scope of the skill beyond what users may expect, increasing the chance of unintended execution paths, accidental file uploads, or invocation of non-audio subcommands under an audio-branded capability.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger keywords are broad, generic phrases like '生成音频', '文本转语音', and '生成音乐, 音效' that can overlap with normal user conversation. In an agent ecosystem that auto-invokes skills based on keyword matching, this can cause unintended activation of a networked CLI that uploads prompts and local media to external services, creating privacy, cost, and workflow-integrity risks.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The trigger keywords are broad and generic, such as 'generate audio' and 'text to speech,' without clear activation boundaries or confirmation requirements. In an agent environment, this can cause the skill to fire on loosely related requests and run external CLI commands or send content to remote APIs when the user did not explicitly intend to invoke this specific third-party tool.

VirusTotal

64/64 vendors flagged this skill as clean.