Security audit
Security checks for vulnerabilities and agentic risk
The skill appears non-malicious, but it should be reviewed because it gives concrete portfolio trading actions without clear financial-risk safeguards.
Install only if you want a Chinese/Beijing-time A-share analysis workflow that fetches public market data and may generate direct-sounding portfolio actions. Treat any add, reduce, exit, switch, or cash-deployment output as analysis to review with your own risk tolerance and professional advice, not as an instruction to trade.
The instruction requires printing 当前北京时间：YYYY-MM-DD HH:MM（UTC+8） in Chinese, and the fixed output contract later mandates Chinese-language labels for key sections. This imposes a specific language on all users without offering a choice or documenting a justified locale restriction.
The activation text is broad enough to trigger this skill for a wide range of stock-analysis requests without clear boundaries or explicit user intent. In a skill that performs online data fetching and produces trading recommendations, overbroad activation increases the chance of unintended invocation, unnecessary exposure of portfolio data, and unrequested financial advice.
The file requires all user time references to be interpreted in Beijing time and requires every answer to present current Beijing time, with no indication that the user can choose another language or locale. This is a natural-language locale policy constraint applied globally rather than as an explicit user opt-in or a clearly justified region-specific tool boundary.
The prompt first mandates loading Chinese fonts only via explicit font file paths and forbids using system/default font families, but later instructs automatic detection/loading of system Chinese fonts such as SimHei/CJK/Heiti. These directives conflict and could cause the skill to behave inconsistently depending on which instruction is followed.
The skill explicitly instructs the agent to generate add/reduce/exit/rotation recommendations for holdings and cash deployment without requiring any risk disclosure, suitability check, or clear limitation that this is not personalized financial advice. In a trading skill, this materially increases the chance of harmful real-world actions by users, especially because the instructions push automatic portfolio decisions rather than descriptive analysis only.
The core task says to execute a continuous workflow of analysis followed by plotting, and the subsequent steps describe mandatory chart generation. However, the final instruction says to skip plotting during intraday trading time, which directly contradicts the earlier unconditional workflow requirement.
The plotting section specifies 'Chinese annotation constraints,' which is a locale/language restriction embedded in the skill instructions. Because the skill does not offer an alternative language or ask for user preference, this can violate language-choice policy.
The default prompt mandates a fixed analysis workflow and output behavior, limiting user control over how the model responds. In a financial-analysis context, this can cause the system to deliver prescriptive trading recommendations or perform networked data retrieval even when the user's intent, locale, or compliance requirements differ.
The skill explicitly instructs online fetching from third-party financial sources and requires audit logging of user-provided symbols and normalized identifiers, but provides no user-facing notice, consent boundary, or retention limits. This creates a privacy and transparency issue because user inputs and trading-related interests may be transmitted to external services and stored in logs without clear disclosure, which is especially sensitive in a finance-related skill.
No suspicious patterns detected.
