Security audit

健康饮食提醒

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Chinese-language wellness reminder skill with no hidden code, exfiltration, or destructive behavior found.

Before installing, confirm you want recurring meal, hydration, exercise, and weekly report reminders, adjust the Asia/Shanghai timezone if needed, and treat the diet guidance as general wellness advice rather than medical advice. Users with pregnancy, breastfeeding, diabetes, eating-disorder risk, or other medical needs should consult a clinician.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Natural-Language Policy Violations

Medium

Confidence: 96% confidence
Finding: The description is written entirely in Chinese and centers the skill around Chinese dietary guidance, but it does not state that the skill is China-specific or offer any language choice. This creates a locale/language policy issue because users are implicitly forced into a specific language and regional standard without opt-in.

Natural-Language Policy Violations

Medium

Confidence: 91% confidence
Finding: The skill configures reminders using the fixed timezone "Asia/Shanghai" for hydration reminders, and the document does not explain that this is limited to that region or allow user choice. Hard-coded locale behavior without opt-in can violate language/locale policy expectations.

Natural-Language Policy Violations

Medium

Confidence: 94% confidence
Finding: All scheduled meal, follow-up, hydration, exercise, and weekly report reminders use the fixed timezone "Asia/Shanghai". Because the skill does not present this as a region-specific tool or ask the user to confirm their locale, it imposes a specific regional setting by default.

Intent-Code Divergence

Medium

Confidence: 91% confidence
Finding: SKILL.md 在目标模式和四季晚餐方案中多次允许‘晚餐可不吃’或‘轻食/不吃’（如 L065、L092、L111、L130、L149），体现的是围绕晚餐选择的饮食建议。而 L231 的晚餐 cron 文案额外加入‘提醒睡前3小时不进食’，这与文件前文实际建议的晚餐时间 18:30 及可不吃晚餐的设定并不一致，属于文档内部对技能意图的主动冲突。

Intent-Code Divergence

Medium

Confidence: 98% confidence
Finding: 文档前文在四季下午茶中明确提供水果、温补、零食等可食用选项，并说明每个饭点从对应池中选 3 个方案 ABC 供选择（如 L087、L106、L125、L144）。但 L233 的 cron 文案要求‘遵循只喝不吃，馋了就喝水’，直接否定了这些已列出的下午茶食物方案，属于明确的意图与实现提示词冲突。

Static analysis

No suspicious patterns detected.