Security audit
Security checks for vulnerabilities and agentic risk
This skill is a prompt-rewriting helper with no hidden access, persistence, or destructive behavior found.
Reasonable to install if you want a Chinese-first prompt optimization helper. Watch for accidental activation from broad wording like "优化一下", and only use it when you actually want the agent to rewrite a prompt.
声明描述的是面向最终用户的 prompt 优化能力，但代码片段只是测试套件，验证技能文档是否满足预期格式和内容要求，例如 frontmatter、RTCF、示例数量、触发词和删除痕迹等。这与“检测提示词缺失要素并输出优化版”的核心功能明显不一致。虽然测试脚本可被视为开发支持文件，但就该 supplied code chunk 本身而言，其实际行为与声明用途不匹配，因此应判定为 mismatch。
The trigger phrases include broad terms such as '优化一下' and 'optimize prompt', which can match ordinary conversation and cause unintended skill activation. Misfiring prompt-rewrite skills can override the user's actual intent, transform benign requests into meta-prompting flows, and increase the chance that untrusted user text is reframed into higher-authority instructions.
This shell file contains natural-language strings in Chinese and labels the script as the practical test version in Chinese, while the test suite later checks for bilingual support. In the context of the policy rule, this indicates a language preference embedded in the skill artifacts without an explicit user choice or opt-in mechanism in this file.
No suspicious patterns detected.
