Security audit

胡开奕 · 养老与传承顾问技能 · 风险评估报告

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently generates local pension and inheritance risk reports, with disclosed sensitive-data handling risks users should manage.

Install only if you are comfortable creating local DOCX/PNG files that may contain sensitive family, health, financial, insurance, property, and inheritance details. Use a private folder, avoid synced/shared directories, redact source notes when possible, and delete or protect generated reports after delivery.

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 (2)

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The report generator writes highly sensitive personal and financial data, including family structure, health status, insurance details, property holdings, and inheritance intentions, to predictable local files (`family_diagram.png` and a `.docx` report) without any safety controls or prominent warning to the operator. In agent or shared workstation environments, these artifacts can persist on disk, be indexed, synced, or accessed by other users or processes, causing confidentiality loss even without a direct code execution flaw.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger list includes broad, generic finance and养老/传承-related phrases such as '风险分析报告', '生成风险评估', and '养老风险盘点', which can match ordinary user requests outside the intended narrow skill scope. This can cause accidental invocation in unrelated contexts, leading the agent to process sensitive family, financial, or inheritance discussions under specialized guidance the user did not explicitly request.

Static analysis

No suspicious patterns detected.