ClawHub
Back to skill

Security audit

口播八步输出模型

Security checks for vulnerabilities and agentic risk

Overview

This is a text-only Chinese short-video script framework with no executable code, persistence, credential access, or hidden data handling.

Before installing, consider whether you want this style of content guidance applied to sensitive social and mental-health-adjacent topics. It appears safe as a text framework, but users should still review outputs carefully for accuracy, stigmatizing language, and platform-sensitive claims.

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The skill description is broadly scoped around multiple loosely defined user intents such as script creation, topic evaluation, optimization, and matrix planning, without clear activation boundaries. This can cause the agent to invoke the skill in situations the user did not intend, increasing the chance of inappropriate behavior, overreach, or policy-bypassing content generation in sensitive domains like workplace manipulation, human nature, and NPD-related analysis.

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
The skill description enforces Chinese-language operation without indicating that the user can choose another language or opt in. This can override user preference, reduce transparency, and lead to unsafe misunderstandings if users receive output in a language they did not request, especially for nuanced or sensitive advice content.

Static analysis

No suspicious patterns detected.