The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script novel_style_runtime_loader.py to process style configuration files. This is a standard utility function for the skill's primary purpose. The script uses json.loads and pathlib for file reading, which is safe for local JSON files.
[DATA_EXPOSURE]: The skill reads local JSON files from a specific directory (novel_styles/). These files are part of the skill's own data structure and do not target sensitive system locations or credentials.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were found. All scripts and prompts are local to the skill package.
[PROMPT_INJECTION]: The instructions contain strict directives ('Hard Rules') to prevent the AI from reproducing copyrighted content or leaking source names, which acts as a safety measure rather than a malicious bypass.
[DYNAMIC_EXECUTION]: The skill uses a simple Python script to format data from JSON into a text-based 'Runtime Card'. It does not use eval(), exec(), or any other unsafe dynamic execution patterns.

use-novel-style