ClawHub
Back to skill

Security audit

山海经童话机

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent Chinese children's story-and-audio skill with expected podcast generation and no evidence of hidden execution, persistence, or malicious behavior.

Install if you want a Chinese-language Shan Hai Jing children's story generator that also creates podcast audio. Avoid putting private, sensitive, or proprietary details into prompts because the resulting story script may be passed to the audio-generation tool.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill instructs the agent to send the complete story script in `user_prompt` to an external `create_podcast` tool, but it does not require notifying the user that their generated content may be transmitted to another service. If the user prompt contains personal, sensitive, or proprietary details, those details could be disclosed to a third-party processor without informed consent.

Vague Triggers

Medium
Confidence
86% confidence
Finding
The skill description is broadly framed as a general story and podcast generator without clear activation boundaries, allowed topics, or safety constraints. In an agent ecosystem, vague scope can cause the skill to be invoked in unintended contexts and increases the chance it will generate age-inappropriate, frightening, or otherwise unsafe content for children.

Natural-Language Policy Violations

Low
Confidence
81% confidence
Finding
The description hard-codes a target audience of '8-year-old girls' and suggests tailoring content to that profile without obtaining user confirmation. This can lead to inappropriate personalization, stereotyping, or mismatched child-directed content when the actual user or listener differs from that assumed audience.

Static analysis

No suspicious patterns detected.