Security audit
Security checks for vulnerabilities and agentic risk
This is a transparent China A-share stock-analysis skill with financial-advice caution points but no hidden code, persistence, credential use, or destructive behavior.
Before installing, treat the output as informational stock research rather than personalized investment advice. Use trusted upstream macro and industry skills, verify current market data independently, and review the final risk section before relying on any buy, hold, or sell suggestion.
The skill explicitly generates buy/hold/sell-style recommendations and star ratings without requiring a financial-risk disclaimer or clarifying that the output is informational rather than personalized investment advice. Because the domain is investing, users may rely on these recommendations for real financial decisions, increasing the risk of harmful overreliance and misinterpretation.
The trigger conditions are broad enough to match many ordinary investment-related prompts, which increases the chance the skill is invoked when the user did not explicitly request this specific workflow. In a financial-advice context, unintended invocation matters because the skill produces structured investment judgments and action recommendations that may overly influence user decisions.
The skill instructs the agent to include the complete outputs of upstream skills verbatim in the final response, which removes any opportunity to filter sensitive, irrelevant, or prompt-injected content from those tools. If an upstream tool returns confidential data, hidden instructions, or malformed content, this skill will propagate it directly to the user.
The file is explicitly written for 'A股个股' analysis and all prompts/output structure are fixed in Chinese, with no user opt-in or alternative language handling. While domain specificity may justify the market scope, the skill text still imposes a specific language/locale experience rather than offering a choice.
No suspicious patterns detected.
