The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to discover an optional companion skill (better-writing) by checking several specific directory paths, including the user's home directory (~/.claude/skills/ and ~/.agents/skills/). It also makes extensive use of a custom story CLI tool and a fallback Node.js script located at a relative path (../story-maintenance/scripts/story.js) to perform chapter indexing, link validation, and continuity checks.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external story project files, creating a surface for indirect instructions that could affect agent behavior during the drafting process.
Ingestion points: Project context is retrieved from files such as story.md, character entities, and plot outlines using the story context command.
Boundary markers: The instructions do not define specific delimiters or instructions to prevent the agent from accidentally executing instructions found within story prose or entity metadata.
Capability inventory: The agent has the capability to write and update markdown files in the chapters/ and scenes/ directories and execute shell commands via the story CLI suite.
Sanitization: There is no mention of sanitization or content filtering for the retrieved story data before it is interpolated into the agent's context for drafting.

chapter-writing