The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses npx markdownlint-cli2 to perform linting and auto-formatting on user-provided Markdown files (SKILL.md). This is a standard utility for document quality control.
[EXTERNAL_DOWNLOADS]: The skill references the markdownlint-cli2 package via npx, which fetches code from the npm registry. This is a well-known service for development tools and is used here for legitimate formatting tasks.
[REMOTE_CODE_EXECUTION]: The skill invokes npx, which downloads and executes the markdownlint-cli2 package from the npm registry at runtime. As the source is a well-known package registry, this is documented as a standard functional behavior.
[PROMPT_INJECTION]: The skill processes user-supplied text for editing and brainstorming, which creates a surface for indirect prompt injection.
Ingestion points: User-provided articles or drafts in Inspiration and Proofreading modes (SKILL.md).
Boundary markers: None identified to separate untrusted user content from agent instructions.
Capability inventory: Command execution via npx (SKILL.md) and file-writing operations for 'material.md' (references/structure-review.md).
Sanitization: No explicit sanitization or filtering of user content before processing is mentioned.

writing-assistant