Security audit

toy-generator

Security checks for vulnerabilities and agentic risk

Overview

This skill is a documentation-only connector for remote toy/3D model generation, with expected external API use and no hidden local persistence or destructive behavior.

Install only if you are comfortable sending prompts, image URLs, and generated toy assets to the OneKey/Craftsman remote service and downstream generation providers. Keep DEEPNLP_ONEKEY_ROUTER_ACCESS and share_url values private, especially links containing pwd or other key-like parameters.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill clearly routes user prompts, image URLs, and generated assets through a remote gateway and third-party generation providers, but it does not explicitly warn users that their content leaves the local agent environment for external processing. This can lead to unintended disclosure of sensitive prompts, proprietary images, or regulated data because users may reasonably assume the skill operates only within the agent platform.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The documentation states that a private share URL includes keys and should be given to the user, but it does not warn that anyone possessing that URL may gain access to task progress, generated images, or model artifacts. Treating bearer-style share links casually increases the chance of accidental disclosure through logs, chat transcripts, screenshots, or copied terminal history.

Static analysis

No suspicious patterns detected.