Image Storyboard

• Outcome: Pass
• Latest audit: Jun 2, 2026 at 10:29 AM UTC
• Version:
  1.2.0

Overview

This skill is a disclosed storyboard and image-generation workflow that uses the dLazy CLI and API key, with no artifact-backed hidden or destructive behavior.

Install only if you are comfortable using the third-party dLazy CLI, storing or passing a dLazy API key, and sending prompts plus any selected local media files to dLazy's cloud services. Prefer npx if you do not want a long-lived global CLI install, and review the CLI source/package before use in sensitive projects.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• 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 (5)

Description-Behavior Mismatch

Medium

• Confidence: 91% confidence
• Finding: The skill’s stated purpose is storyboard planning with gated progression, but the later section expands behavior into a general terminal-executing image-generation agent. This broadens operational scope beyond what the manifest and earlier workflow imply, increasing the chance an agent will run commands or perform actions the user did not reasonably expect from a planning-oriented storyboard skill.

Context-Inappropriate Capability

High

• Confidence: 96% confidence
• Finding: The skill explicitly instructs the agent to execute terminal commands using an external CLI, including handling authentication and media uploads to remote endpoints. For a storyboard skill, this introduces powerful side effects—code/package execution, network transmission of prompts and local files, and dependency trust on a third-party tool—that are not inherently required for planning and can expose sensitive local data or expand the attack surface.

Intent-Code Divergence

Medium

• Confidence: 83% confidence
• Finding: The file advertises a strict 'plan first, render later' workflow, but later instructions permit generation after only first-image prompt confirmation rather than full script and storyboard approval. This inconsistency can bypass intended review gates, causing premature execution, unwanted uploads, or generation before requirements are fully validated.

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The skill’s declared purpose is a constrained storyboard workflow, but the embedded instructions expand behavior into a general-purpose terminal-driven image generation agent. That mismatch weakens least-privilege expectations and can cause the host agent to execute commands beyond what a user reasonably infers from the manifest, especially because the skill explicitly frames the agent as able to run terminal commands.

Context-Inappropriate Capability

Medium

• Confidence: 96% confidence
• Finding: The skill explicitly requires terminal command execution and instructs the agent to run CLI commands, which grants a stronger capability than is justified by a storyboard planning skill. In context, this is more dangerous because the skill also handles local file paths and remote uploads, so terminal access can become a bridge to unintended local data exposure, arbitrary command execution patterns, or misuse of installed tooling.

VirusTotal

64/64 vendors flagged this skill as clean.