Security audit
Security checks for vulnerabilities and agentic risk
This is a Markdown-only market research framework that is broadly disclosed, but users should know it defaults to Chinese output and a United States market lens.
Install this only if you want a structured market research assistant that normally writes in Chinese and analyzes from a US-market perspective. For non-US markets or another output language, explicitly override those assumptions in your prompt and verify business-critical recommendations with current sources.
Forcing all user insights and recommendations to a fixed US-market perspective without user choice can systematically distort outputs when the user asks about other regions. In this skill context, that creates a high risk of materially incorrect market advice, false assumptions about users, and decisions based on the wrong regulatory, cultural, and channel environment.
The pre-analysis workflow hard-codes the target market to the United States for every task, which can override clear user intent and contaminate all downstream modules with the wrong locale assumptions. Because the skill produces decision-oriented research reports, this increases the likelihood of harmful business recommendations, invalid competitor sets, and inaccurate consumer behavior analysis.
The trigger list is broad and includes common phrases like market research, product analysis, and user insights, which can cause the skill to activate for ordinary requests that may not need this rigid workflow. Over-broad activation increases the chance of unintended instruction takeover, where the skill's fixed constraints override user intent and lead to irrelevant or misleading outputs.
The activation section defines many positive matches but no boundaries for when the skill should not load, so unrelated or partially related requests may be captured. In an agent setting, this can create prompt-scope confusion and force the assistant into a prescriptive analysis mode that suppresses safer or more appropriate handling of the user's actual request.
The instruction “默认中文输出” imposes a specific output language as a fixed rule, but the document does not offer the user a language choice or opt-in mechanism. This is a natural-language policy concern because it forces locale behavior regardless of user preference.
The document states '输出默认使用中文' as a fixed rule, which imposes a specific language on outputs. The policy allows locale or language constraints only when the user is given a choice or when the constraint is clearly documented and justified as region-specific; neither condition is met for output language here.
No suspicious patterns detected.
