The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill facilitates the installation and execution of remote code from external sources. This includes using the shadcn CLI to add vendor-specific components (@snapcn/*) and employing npx skills to fetch and integrate a secondary 'umbrella' skill from a remote repository (snapcndev/snapcn). Additionally, it provides instructions to add a Model Context Protocol (MCP) server directly from a vendor package (@snapcn/mcp).
[COMMAND_EXECUTION]: Multiple shell commands are utilized to manage the development and production workflow. These include installing UI components via npx shadcn@latest add, generating still frame previews with npx remotion still, and performing full high-definition video renders using npx remotion render with specific codec configurations.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied text through various component properties, creating a surface for indirect prompt injection.
Ingestion points: User-provided strings are passed into properties such as text, words, script, lead, and morphTo within SKILL.md.
Boundary markers: There are no explicit boundary markers or instructions to treat the interpolated text as untrusted data.
Capability inventory: The skill has the capability to execute shell commands (npx shadcn, npx remotion) as documented in SKILL.md.
Sanitization: There is no evidence of sanitization or validation of the input text before it is used to generate code or command-line arguments.

remotion-kinetic-typography