The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingest untrusted data from external text files (novels) and processes them to generate summaries and extract assets.
Ingestion points: The read_novel_segment function in read_novel.py reads content from a user-provided file path.
Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to ignore embedded commands within the novel text.
Capability inventory: The skill allows reading local files via read_novel.py and instructs the agent to write several files (大纲.txt, 读取进度.txt, and multiple character/item/scene files) based on the analyzed content.
Sanitization: No sanitization or filtering of the novel's text is performed before it is passed to the agent for analysis.
[RESOURCE_EXHAUSTION]: The script read_novel.py reads the entire content of the file into memory using f.read() in both get_novel_info and read_novel_segment functions. This approach is inefficient for a tool designed for 'long text novels' and could lead to memory exhaustion (Denial of Service) if executed on exceptionally large files.

novel-reader