Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only Chinese strategic-thinking advice skill with no code execution, persistence, credential use, or data access, though its activation language is somewhat broad.
Install this if you want a Chinese-language Zhang Yiming/ByteDance-style decision framework for strategy, startups, product, or management questions. Treat its quotes, 2026 market/policy notes, and investment-adjacent examples as advisory framing rather than verified facts or financial advice, and invoke it explicitly for best results.
The "When to Use" section is ambiguous and insufficiently bounded, so the runtime may activate the skill for a wide set of generic advice-seeking prompts. This is dangerous in context because the skill is not a neutral general assistant; it applies a specific named thinker’s lens, which can override user intent and produce inappropriate guidance in unrelated situations.
The activation example includes a very broad phrase such as "要不要做XXX？", which can match many ordinary user questions outside this skill’s intended domain. In an agent environment, this can cause unintended invocation and steer unrelated conversations into this persona/framework, leading to prompt hijacking of routing behavior and reduced reliability of other skills.
Declaring the skill as "中文为主，英文关键术语保留" imposes a language preference without indicating user override or consent. In a multi-user or multilingual agent setting, this can degrade usability, cause misunderstanding, and create unintended behavior if the skill activates for users who did not request Chinese output.
