Security audit
Security checks for vulnerabilities and agentic risk
This is a Markdown-only Chinese film-commentary writing aid with no code execution, persistence, credential access, or hidden data handling.
Installers should understand this skill is intended for Chinese-language short-video film commentary. It may activate on broad media-production terms, so users should be explicit when they want a different kind of BGM, subtitle, or editing task. Also review music and film-clip copyright before publishing outputs.
The skill metadata and content are written as Chinese-only and do not offer a language fallback or user-choice mechanism, which can force the agent into a language the user did not request. While not a direct security exploit, this is a policy and routing weakness because it can degrade user control and cause unintended behavior in multilingual environments.
The trigger conditions are broad enough to activate on generic phrases like '解说文案', '分段字幕', 'BGM 推荐', or any movie name plus duration/style combination, which can cause the skill to hijack unrelated requests. This creates prompt-routing confusion and may override user intent, reducing reliability and potentially causing the agent to disclose or generate content in the wrong context.
The entire skill file is written as a prescriptive reference in Chinese and does not indicate that users may choose another language or locale. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly justified.
