ClawHub
Back to skill

Security audit

Self Learning Coach Deep v0.1.2

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed learning-coach workflow that creates lesson files and uses user-provided or web sources, with no evidence of hidden exfiltration, destructive actions, or excessive privilege.

Install this if you want an agent to create structured learning lessons and local progress/source records. Be aware that it may read user-provided Feishu or local materials through read-only tools, write lesson files under lessons/, and send generated HTML through Feishu when used in that environment; avoid providing sensitive documents unless you are comfortable with those lesson and source records being created.

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 (3)

Missing User Warnings

Low
Confidence
77% confidence
Finding
The skill instructs the agent to create and update HTML and tracking files in the workspace by default, but it does not consistently require explicit user notice or consent before writing artifacts. In environments where workspace contents are sensitive, this can lead to unintended persistence of user data, source metadata, or internal learning records beyond the user's expectations.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The default invocation phrase is broad enough to activate the skill for many generic learning requests without clear user intent boundaries. This can cause the agent to route ordinary queries into a specialized workflow unexpectedly, increasing the chance of unintended file creation, external-source use, or overbroad task execution in contexts where the user did not explicitly request this skill.

Natural-Language Policy Violations

Medium
Confidence
82% confidence
Finding
The skill metadata and default prompt are written in Chinese and appear to steer the interaction into Chinese without checking the user's preferred language. While not directly enabling code execution or data exfiltration, this can override user expectations, reduce transparency, and cause misunderstandings in security-relevant or business-critical learning flows.

Static analysis

No suspicious patterns detected.