The Agent Skills Directory

[SAFE]: The skill is designed for creative writing assistance, specifically for maintaining voice and style consistency across a manuscript. It operates entirely on local project files such as story.md and style-sheet.md.
[SAFE]: Command execution is limited to the 'story' CLI tool suite (e.g., story prose, story voices, story validate). These tools are executed locally via bun or node from a relative path (../story-maintenance/), representing standard local development tool integration.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection as it processes user-provided manuscript text and style guides which could contain instructions intended to influence the agent.
Ingestion points: The skill reads story.md, style-sheet.md, and individual chapter files during the analysis and drafting workflow (SKILL.md, Workflow Step 1).
Boundary markers: None are explicitly mentioned for delimiting user prose from instructions.
Capability inventory: The skill uses the story CLI to analyze text metrics and update word counts (SKILL.md, Act on the findings).
Sanitization: No specific sanitization or filtering of the ingested manuscript text is described.
[SAFE]: No evidence was found of data exfiltration, credential harvesting, obfuscation, or persistence mechanisms. The skill focuses strictly on linguistic analysis and project maintenance tasks.

voice-style