The Agent Skills Directory

[SAFE]: The skill provides a logical framework for organizing creative writing projects using plain text files within the local workspace. It does not attempt to access sensitive system areas or make network connections.
[COMMAND_EXECUTION]: The instructions suggest using the grep utility to search for specific terms within a dedicated sub-directory (设定/). This is a legitimate use of a common tool for content management within the scope of the novel project.
[PROMPT_INJECTION]: While the skill uses the term "injected" in its header, the context is purely functional, indicating the activation of the skill's logic rather than an attempt to override the underlying safety protocols or system prompts of the AI.
[DATA_EXPOSURE]: The skill operates on user-generated content (novel drafts and metadata) and maintains all data within a specific project folder, presenting no risk of exposing sensitive credentials or system files.

novel-memory