Security audit

智能保洁助手

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent local cleaning-management skill, but users should know it may save cleaning activity from broad natural-language triggers and its HTML reports load Chart.js from a CDN.

Install only if you are comfortable with a local SQLite cleaning log being created under the skill directory. Be explicit when asking it to save records, review or delete mistaken entries, and avoid opening generated reports in sensitive contexts unless you accept the CDN dependency for Chart.js.

SkillSpector

Vulnerability Patterns

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
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: The generated report loads Chart.js from a public CDN, which causes the local report to execute third-party JavaScript whenever it is opened. If the CDN content is compromised, replaced, or blocked, users can be exposed to supply-chain risk, tracking, or report malfunction despite the skill being described as local and SQLite-based.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The top-level trigger list contains broad everyday phrases such as '家务' and '今天打扫' that can match normal conversation and cause unintended activation. Because the skill has Write, Bash, and WebFetch tools and is designed to persist records locally, accidental invocation could lead to unwanted data creation or command execution paths without the user explicitly intending to use the skill.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The module trigger phrases for planning include ambiguous language like '生成计划' and '今天打扫什么', which could be interpreted from ordinary user requests outside an explicit skill context. In this skill, unintended activation is more concerning because it can invoke local scripts and potentially initialize or modify SQLite data automatically.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The natural-language recording triggers are broad enough to capture casual statements about cleaning activity, not just requests to save a record. Since the skill then parses details and stores them in a local database, this can create privacy and integrity issues through accidental record creation from ordinary chat.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The skill explicitly states that user descriptions of cleaning activity are automatically parsed and saved, but it does not require a warning or confirmation before writing to storage. This is dangerous because users may disclose routines, dates, locations, and habits in normal conversation, and the skill would persist that data locally without clear consent, creating privacy risks and inaccurate records.

Static analysis

No suspicious patterns detected.