The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill contains instructions to download the Typst binary directly from the official Typst project releases on GitHub (https://github.com/typst/typst/releases). This is a well-known open-source tool required for the skill's primary functionality.
[COMMAND_EXECUTION]: The skill executes shell commands to verify the installation of Typst (typst --version) and to compile document templates into image files (typst compile).
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it is designed to ingest and act upon content from untrusted external files such as DESIGN.md and brand logo files.
Ingestion points: The agent is instructed to read DESIGN.md and logo files (PNG, SVG) provided by the user to extract design tokens (colors, fonts, styles).
Boundary markers: There are no explicit instructions or delimiters provided to the agent to distinguish between design instructions and potential malicious commands embedded within the user files.
Capability inventory: The skill uses subprocess calls to execute the typst binary for rendering.
Sanitization: No specific sanitization or validation steps are described for the content read from user-provided documents before it is used in the rendering process.

typst-cards