ClawHub
Back to skill

Security audit

图表制作大师(专业版)

Security checks for vulnerabilities and agentic risk

Overview

This diagram skill is mostly purpose-aligned, but it under-discloses network/API behavior while asking for read, write, and command execution authority.

Install only if you are comfortable letting the agent read diagram/document inputs, write output files, and run local diagram/export commands. Avoid sensitive documents unless you understand the agent platform's LLM data handling, disable or review callback URLs and external API use, and treat interactive SVGs as active content from trusted sources only.

SkillSpector

Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (5)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill claims all generation/export happens locally and that no data is uploaded, yet it also documents online-loaded Google Fonts and dependence on an external LLM API provided by the agent platform. This creates a misleading privacy guarantee that could cause users to expose sensitive document content under false assumptions about data locality.

Context-Inappropriate Capability

Medium
Confidence
84% confidence
Finding
The skill advertises generic API integration and external service calling even though its stated purpose is local diagram generation. Broad API capability increases the risk of unintended network access, data exfiltration, or scope creep if the agent interprets the skill as permission to contact arbitrary services.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The activation text is overly broad and ambiguous, covering general design and branding scenarios beyond diagram generation. This can cause the skill to trigger in unrelated contexts and gain access to read/write/exec capabilities when the user did not intend to invoke a command-capable automation tool.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The documentation describes save-version, rollback, and other modifying operations without a prominent warning that local files and history can be changed. Users may unintentionally overwrite diagrams or revert content, which is especially risky in collaborative or automated workflows.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The interactive SVG feature supports embedded links and JavaScript-capable viewers but omits a strong warning about privacy and active content risks. Malicious or careless link embedding could create phishing, tracking, or script-execution exposure when generated SVGs are opened in capable environments.

Static analysis

No suspicious patterns detected.