The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches components from the external @elements registry via the shadcn CLI during execution.
[COMMAND_EXECUTION]: Instructs the agent to run shell commands to add components (npx shadcn@latest add) and inspect local registry files (ls).
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the interpolation of user-supplied variables into shell commands.
Ingestion points: {name} placeholder in SKILL.md used for logo selection.
Boundary markers: No delimiters or protective instructions are used to wrap the interpolated input.
Capability inventory: Access to shell execution through npx and directory listing as shown in SKILL.md.
Sanitization: There is no validation or escaping of the {name} input prior to its use in the shell command string.

tech-logos