The Agent Skills Directory

[COMMAND_EXECUTION]: The skill launches a local Node.js server using scripts/dashboard-server.mjs to provide the 'Dashboard' web interface for editing novel files. It also uses the spawn module to automatically open the local dashboard URL in the user's default web browser.
[COMMAND_EXECUTION]: A Python script (scripts/next_step.py) is executed to analyze the project's file structure and determine the current writing progress, which helps route the user to the appropriate next step in their workflow.
[EXTERNAL_DOWNLOADS]: The skill performs a version check by fetching release metadata from the author's GitHub repository via the https://api.github.com/repos/Chained1001/mo-shu/releases/latest endpoint. Actual updates are only performed if the user manually confirms the action.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and route novel-related data (markdown,大纲, settings) which are externally managed files. While this creates a surface for indirect prompt injection where the agent might follow instructions embedded in a novel draft, the skill implements standard controls and the risk is consistent with its primary function as a text editor and router.

moshu