ClawHub
Back to skill

Security audit

图表生成工具-专业版

Security checks for vulnerabilities and agentic risk

Overview

This diagram-generation skill mostly matches its purpose, but it asks for broad local execution and file access while leaving callback data sharing, first-run filesystem changes, and trigger scope under-scoped.

Review this before installing if you use sensitive architecture or business diagrams. Use it only for explicit diagram-generation tasks, inspect the npx adapter package before enabling it, avoid callback_url unless you control the HTTPS endpoint, and choose or confirm output paths before first run.

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 (4)

Vague Triggers

High
Confidence
86% confidence
Finding
The trigger scope is overly broad, including generic productivity and workflow-optimization language that can cause the skill to activate for unrelated user requests. In a skill with read/exec/write capabilities, overbroad routing increases the chance of unintended command execution, file creation, or tool use without informed user intent.

Vague Triggers

High
Confidence
89% confidence
Finding
The keyword coverage list contains vague everyday-language terms such as 'Use', 'when', and broad workflow phrases, making accidental invocation likely. Because the skill advertises automation plus local command/file capabilities, imprecise matching can expand operational reach beyond the user's actual request and lead to unintended side effects.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The documentation states that first use will automatically create local configuration files and output directories, but does not surface this prominently as a user-facing warning before initialization. Silent filesystem modification is risky in agent contexts because users may not expect persistent changes or know where data is being written.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
Documenting callback_url support without warning that task content may be transmitted to a remote endpoint creates a real privacy and data-handling risk. Users may provide sensitive diagram specs, architecture data, or workflow details that could be exfiltrated over the network if callbacks are enabled without clear consent and safeguards.

Static analysis

No suspicious patterns detected.