Security audit

portfolio management

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Chinese-language portfolio analysis skill that reads user-provided holdings screenshots and produces market research, with no code execution, persistence, or hidden installation behavior found.

Install only if you want a Chinese-language portfolio review workflow. Before uploading screenshots, redact account numbers, broker identifiers, personal details, and any fields you do not want analyzed. Treat recommendations as research support, not a substitute for licensed financial advice or your own risk judgment.

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

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (7)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The skill description uses very broad trigger phrases such as stock analysis, position analysis, investment advice, pre-market analysis, post-market review, or simply uploading a holdings screenshot. This can cause unintended invocation in loosely related conversations, leading the agent to process sensitive financial data or provide regulated advice when the user did not clearly request this skill.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The workflow instructs the agent to read user-uploaded holdings screenshots and then perform external searches, but it does not warn about privacy risks or limit what sensitive data may be extracted and transmitted. Holdings screenshots can reveal account positions, cost basis, profits/losses, and allocation data, which are highly sensitive financial information and may be unnecessarily exposed through downstream processing or third-party services.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The framework gives explicit buy/hold/reduce/liquidate recommendations and position-sizing guidance without any financial-risk warning, suitability check, or limitation statement. In an investment-advice skill, this can cause users to treat the output as personalized financial advice and act on it despite missing disclosures about market risk, uncertainty, and the need for independent judgment.

Natural-Language Policy Violations

Medium

Confidence: 91% confidence
Finding: The file title and all operational instructions are written as mandatory Chinese guidance, and the template/output examples are exclusively in Chinese. There is no indication that users may choose another language or locale, which can violate a language/locale policy when user opt-in is required.

Natural-Language Policy Violations

Medium

Confidence: 94% confidence
Finding: The entire template is written as a fixed Chinese report format, including headings, labels, and reminders, with no indication that another language can be used. This is a natural-language policy concern because it imposes a specific language/locale by default rather than documenting user choice or justified regional scope.

Natural-Language Policy Violations

Medium

Confidence: 96% confidence
Finding: 该技能文档从标题开始即以中文呈现全部操作说明，未向用户提供语言/locale 选择，也未说明这是仅面向特定中文用户群或地区的限定技能。根据语言/locale 政策，强制单一语言而无用户选择属于自然语言层面的合规问题。

Natural-Language Policy Violations

Low

Confidence: 86% confidence
Finding: 整份技能说明以中文固定描述功能、流程和输出模板，未说明是否支持其他语言，也未提供用户选择语言的入口。根据语言/locale 政策，若技能默认强制单一语言而无用户选择或明确的地区性理由，属于自然语言策略约束问题。

Static analysis

No suspicious patterns detected.