ClawHub
Back to skill

Security audit

字体管理专业版

Security checks for vulnerabilities and agentic risk

Overview

This looks like a font-management skill, but its trigger and capability claims are broad enough that users should review it before installation.

Install only if you intend to use it for font and typography management. Before use, narrow the trigger to typography tasks, confirm any local Python command before execution, review output paths for generated CSS/fonts/reports, and avoid providing CDN API keys or callback URLs unless the workflow truly needs them.

SkillSpector

Vulnerability Patterns
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The skill is presented as a font-typography governance tool, but later claims generic capabilities such as API integration, command execution, file handling, and information retrieval. This scope expansion can cause an agent or user to authorize actions far beyond the expected domain, increasing the chance of unnecessary file access, network calls, or command execution under a misleading trust boundary.

Intent-Code Divergence

Medium
Confidence
92% confidence
Finding
The trigger-condition text instructs use for translation, multilingual conversion, and localization, which contradicts the rest of the file describing a font-management utility. Such mismatch can route unrelated user requests into a skill with exec/write capabilities, causing the agent to perform inappropriate actions in contexts where the user did not intend to invoke a file- or command-capable tool.

Vague Triggers

High
Confidence
90% confidence
Finding
An overly broad and mismatched trigger condition is dangerous because it can cause automatic or heuristic selection of this skill for unrelated tasks. In this case, the risk is amplified by the presence of read/exec/write tools, meaning a misrouted invocation could lead to local modifications or command execution under false pretenses.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill describes commands and outputs that write files or generate artifacts, but does not prominently warn that local files may be modified or created. In an agent setting, this weakens informed consent and can lead users to invoke the skill without understanding that CSS, report directories, or generated font assets may be changed on disk.

Static analysis

No suspicious patterns detected.