The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The script scripts/shadcn_add.py manages component installation securely.
It utilizes subprocess.run() with a list of arguments, which prevents shell injection attacks by ensuring input is not interpreted by a command shell.
User input for component names is treated as discrete arguments to the shadcn CLI.
[EXTERNAL_DOWNLOADS] (SAFE): The skill documentation identifies standard package management workflows.
It uses npx to fetch the latest shadcn CLI from the official npm registry.
External links for documentation (e.g., tailwindcss.com, ui.shadcn.com) are trusted sources.
[REMOTE_CODE_EXECUTION] (SAFE): Remote execution occurs through standard developer tools (npx).
The execution is targeted at well-known, community-standard libraries.
The Python helper script uses best practices for invoking these external tools.
[DATA_EXFILTRATION] (SAFE): No evidence of unauthorized data access or exfiltration was found.
Scripts do not access sensitive file paths like SSH keys or environment secrets.
[PROMPT_INJECTION] (SAFE): The skill metadata and documentation are free of instructions aimed at overriding agent behavior or bypassing safety filters.

ui-styling