The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses dynamic context injection in SKILL.md via the !npx shadcn@latest info --json`` command. This results in the automatic execution of a shell command when the skill is loaded to provide the agent with project configuration data.
[REMOTE_CODE_EXECUTION]: The add command documented in cli.md allows for the installation of components from arbitrary external URLs (e.g., npx shadcn@latest add https://api.npoint.io/abc123). This capability can be exploited to download and execute untrusted code if the source is malicious.
[EXTERNAL_DOWNLOADS]: The skill workflow in SKILL.md and cli.md involves fetching content from external URLs, such as component documentation and usage examples generated by the npx shadcn@latest docs command.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its processing of untrusted external data.
Ingestion points: Documentation and example URLs fetched from the docs command, and registry item content viewed via the view command.
Boundary markers: No boundary markers or 'ignore' instructions are provided for the ingested data.
Capability inventory: High-privilege shell command execution via the npx shadcn@latest add tool.
Sanitization: No sanitization or validation of the remote content is performed before processing.

shadcn