The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to execute @mizchi/vlmkit-anim using npx, which downloads and runs the animation tool. This is a vendor-owned resource belonging to the skill author.
[COMMAND_EXECUTION]: The skill utilizes several CLI commands for checking scene validity, analyzing layout, and rendering figures (e.g., vlmkit-anim check, vlmkit-anim facts, vlmkit-anim repo). These commands operate on local files to generate documentation artifacts like SVGs, GIFs, and HTML files.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect injection as it ingests and processes untrusted data from external sources such as repository source code, git branch history, and existing Markdown files containing Mermaid diagrams.
Ingestion points: Directory paths, git branch references, and Markdown files passed to the facts, repo, pr, and import commands in SKILL.md.
Boundary markers: None explicitly specified for the ingestion of external codebase or PR data.
Capability inventory: The skill executes subprocesses (via npx or node), reads local filesystem data (code, git history, markdown), and writes output files (SVG, PNG, HTML, GIF).
Sanitization: The vlmkit-anim check command performs schema validation and semantic checks on the generated JSON scenes before rendering.

explanatory-animation