The Agent Skills Directory

[SAFE]: No security issues were detected. The skill acts as a documentation resource for the Takumi rendering library.
[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the takumi-js library and its associated packages from standard registries using package managers like bun. It also describes fetching font resources from well-known services such as Google Fonts and Bunny Fonts for typography rendering.
[INDIRECT_PROMPT_INJECTION]: The skill provides an interface for rendering JSX and HTML content, which constitutes an attack surface if untrusted user data is processed. However, the documentation proactively addresses these risks:
Ingestion points: The render and ImageResponse functions accept JSX, HTML strings, or node trees as input for rendering (found in SKILL.md and references/setup-core.md).
Boundary markers: The documentation recommends using the images.allowUrl option to create an allowlist for remote resources, preventing unrestricted fetching (found in SKILL.md).
Capability inventory: The skill facilitates rendering nodes to various image formats (PNG, JPEG, WebP) and SVG, which involves fetching remote fonts and images (found in references/fonts-images-styling.md).
Sanitization: The skill explicitly warns about SSRF risks and advises on the use of fetch timeouts and isolated caches to mitigate these concerns (found in SKILL.md and references/pitfalls-migration.md).

takumi