Security audit

learning-dashboard-plugin

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small instruction-only learning tracker that edits two named local markdown files, with the main risk being accidental updates from casual trigger phrases.

Install this only if you want your agent to update wiki/projects/错题集.md and wiki/projects/AI学习全系列课程地图.md when saving or ending a lesson. For short phrases like "存档" or "下课", ask the agent to confirm the files and planned changes before writing.

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger list includes broad everyday phrases such as “存档”, “下课”, and “今天到这”, which can appear in normal conversation without a clear intent to modify files. Because this skill performs writes to project documents, accidental activation could cause unauthorized or unintended state changes to learning records.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill directs the agent to modify specific project files but does not require explicit user awareness or confirmation that persistent files will be changed. In a conversational setting, this increases the risk of silent writes to user data, especially when combined with ambiguous triggers and default behavior to update everything.

Vague Triggers

Low

Confidence: 85% confidence
Finding: The description says the skill should run after finishing a lesson and when the user says broad phrases like "更新进度", "存档", or "下课", but it does not define stronger boundaries such as required learning context, confirmation, or specific state checks. This can cause unintended activation and accidental modification of progress or mistake-book data, which is a genuine security and integrity concern even if not overtly malicious.

Static analysis

No suspicious patterns detected.