The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a local Python script scripts/check_chapter_wordcount.py to automate word count verification for generated chapters. The script is called via standard command-line execution and performs analysis on the content of the chapter files within the project directory.
[INDIRECT_PROMPT_INJECTION]: The agent reads data from the novel's outline and previously written chapters to maintain narrative consistency and track character states.
Ingestion points: Files located in the novels/ directory, specifically 00-大纲.md and individual chapter files.
Boundary markers: The skill does not define specific delimiters for separating user-controlled content from instructions, though the data being read is agent-generated creative fiction.
Capability inventory: The skill possesses file read/write capabilities and the ability to execute a local Python script.
Sanitization: No specific sanitization or filtering is applied to the creative content before it is read back into the context.
[SAFE]: The skill demonstrates safe software practices by providing a dedicated secret-free environment. There is no evidence of network activity, credential exposure, or persistence mechanisms. All external resources (templates and guides) are included locally within the skill package.

chinese-novelist