Security audit

生活满意度与平衡优化

Security checks for vulnerabilities and agentic risk

Overview

The skill appears purpose-built for life-balance reports, but it should be reviewed because sensitive wellness notes are saved in the browser while third-party JavaScript is loaded.

Install only if you are comfortable entering life, burnout, finance, and reflection data into a local HTML report that saves browser history. Prefer a version that bundles Chart.js locally, asks before saving notes, and provides a clear delete-history control.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (6)

Description-Behavior Mismatch

Medium

Confidence: 95% confidence
Finding: The template loads Chart.js from a third-party CDN at runtime, which introduces external network dependency and a supply-chain trust risk. If the CDN response is tampered with, unavailable, or swapped for a malicious script, the page could execute attacker-controlled JavaScript in the context of a report containing personal self-assessment data.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: The page stores personal scores and free-form notes in localStorage, which persists sensitive wellness and reflection data beyond the active session. Any script running in the same origin, including compromised third-party code, can read this data, and users are not given a clear choice about retention.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger phrases are broad, generic life-coaching terms such as '自我提升', '生活评估', and 'burnout check' that can easily appear in ordinary conversation. This can cause unintended invocation of the skill in unrelated contexts, leading to unexpected collection or processing of sensitive personal wellbeing data and confusing agent behavior.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The skill metadata includes a long list of generic trigger phrases such as '自我提升', '习惯调整', and 'burnout check' that are broad enough to match many ordinary wellness or coaching requests. This can cause unintended activation outside the user's actual intent, leading the agent to steer conversations into structured self-assessment and report generation when a narrower response would be more appropriate.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The activation section treats vague emotional-state phrases like '最近很累', '感觉倦怠', and '帮我调整习惯' as sufficient to launch the skill. Because these phrases commonly appear in unrelated contexts, the skill may over-trigger and begin collecting sensitive wellness information or generating a persistent HTML report without clear user intent for this workflow.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: The application stores sensitive self-assessment notes and scores in localStorage without warning the user that intimate mental-health, burnout, and life-balance information will be retained in the browser. In this skill context, the content is more sensitive than ordinary preferences, so silent persistence increases privacy harm if the device is shared or the origin is later exposed to malicious script.

Static analysis

No suspicious patterns detected.