Security audit

胡开奕 · 养老与传承顾问技能 · 解决方案

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent local document-generation skill for retirement and inheritance planning, with sensitive inputs expected for its purpose and no evidence of hidden collection, network use, or destructive behavior.

Use this only in an approved environment for client financial and health information. Redact unnecessary identifiers before pasting source materials, review the generated plan for compliance before sharing, and store or rename the generated DOCX carefully because it may contain sensitive client details.

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)

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The README explicitly instructs users to paste a risk assessment report and follow-up communication records, which are very likely to contain highly sensitive personal, financial, health, and family information. Without any privacy warning, minimization guidance, redaction requirement, or storage/handling notice, users may disclose regulated or confidential data into the skill workflow in an unsafe or noncompliant manner.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The script writes a DOCX containing sensitive customer financial and health information to a predictable local path without any overwrite confirmation, file-permission hardening, or privacy handling guidance. In this skill context, the generated document is explicitly populated with personal, medical, inheritance, and insurance-planning data, so accidental overwrite, unintended local exposure, or mishandling can directly leak highly sensitive information.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger list includes broad phrases such as “养老解决方案”, “养老金方案”, and “传承解决方案” that can match ordinary user requests for retirement or legacy planning, causing the skill to activate outside its intended narrow context. In a wealth-management/insurance workflow, unintended activation can steer users into regulated financial-planning content without clear intent, increasing the risk of inappropriate guidance, privacy over-collection, or policy bypass through trigger collision.

Static analysis

No suspicious patterns detected.