The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (SAFE): The skill recommends using npx shadcn@latest for initialization and adding components. This is the standard distribution method for this tool.
[COMMAND_EXECUTION] (SAFE): Installation commands are generated by MCP tools and executed in the shell, which is necessary for the skill's functionality.
[PROMPT_INJECTION] (SAFE): No direct injection or bypass patterns were found.
[DATA_EXPOSURE] (SAFE): No sensitive file access or data exfiltration detected.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill processes data from external UI registries. This is a potential attack surface but inherent to the skill's utility.
Ingestion points: shadcn___view_items_in_registries and shadcn___get_item_examples_from_registries in SKILL.md and references/workflows.md.
Boundary markers: Absent for registry content.
Capability inventory: Shell command execution for installation.
Sanitization: Relies on MCP tool validation.

shadcn-management