ClawHub
Back to skill

Security audit

个性化学习助手免费版

Security checks for vulnerabilities and agentic risk

Overview

The skill is a basic learning assistant, but it asks for broad read, write, and command execution authority without explaining why that level of access is needed.

Review this skill before installing. It does not show malicious behavior, but it gives a tutoring assistant command execution and file read/write authority without clear limits. Use it only in a controlled workspace and avoid passing sensitive files, credentials, or private study material to callback-based workflows unless you understand where the callback data goes.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
The skill requests read, write, and exec capabilities even though its documented purpose is a basic learning assistant for concept explanations and study guides. This creates unnecessary authority that could be abused to read local files, modify user data, or execute commands if the skill is invoked in an unexpected context or prompt-injected through user input.

Vague Triggers

Medium
Confidence
72% confidence
Finding
The description claims broad applicability across vague 'development scenarios' and reusable workflows without clear activation boundaries. In agent environments, this can cause the skill to be selected or trusted in contexts far beyond harmless tutoring, increasing the chance that its elevated tools are used on sensitive tasks.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill documents a callback_url for asynchronous completion but does not warn that this may send data to an external destination. In practice, this can enable silent outbound transmission of user prompts, generated content, metadata, or internal results to attacker-controlled endpoints, especially dangerous in an agent with file and exec capabilities.

Static analysis

No suspicious patterns detected.