ClawHub
Back to skill

Security audit

艺术创作助手(专业版)

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly an art-coaching prompt, but it asks for broad read/write/exec authority and includes conflicting data-flow and invocation instructions.

Review before installing. This does not show clear malicious code, but the skill asks for powerful local capabilities and gives inconsistent privacy and network statements. Install only if you are comfortable with an art assistant that may be granted file, command, and possible external-service authority, or require the publisher to narrow the tools and clarify data handling first.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Intent-Code Divergence

Medium
Confidence
92% confidence
Finding
The documentation makes a strong privacy claim that no data is uploaded and portfolio evaluation is fully local, but elsewhere advertises external API integration and network communication. This contradiction can mislead users into sharing artworks or related data under false assumptions, creating privacy and trust risks if the skill or hosting agent transmits content externally.

Context-Inappropriate Capability

Medium
Confidence
90% confidence
Finding
The skill describes broad file-processing, API-integration, and command-execution abilities that are not necessary for an art-coaching skill and are presented at a generic capability level. Overbroad capabilities increase attack surface and can normalize unsafe use of powerful tools for unrelated tasks, especially when paired with read/write/exec permissions in the metadata.

Vague Triggers

High
Confidence
96% confidence
Finding
The invocation guidance says to use the skill for data analysis, report generation, statistical insights, and visualization, which is materially misaligned with the rest of the document describing an art-creation coach. Such vague and conflicting routing criteria can cause the agent to invoke this skill in inappropriate contexts, potentially exposing unrelated user data to read/write/exec-capable tooling.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill advertises command execution, file writing, and callback/external-service style behavior, but the quick-start and user-facing sections do not clearly warn that these operations may occur. Users may reasonably assume they are interacting with a purely conversational art assistant, leading to uninformed consent around filesystem or network-affecting actions.

Static analysis

No suspicious patterns detected.