The Agent Skills Directory

[COMMAND_EXECUTION]: Executes the zard-cli tool via package runners (npx, pnpm dlx, bunx) to automate the addition and overwriting of UI components. The allowed-tools frontmatter explicitly scopes shell access to these specific command patterns.
[EXTERNAL_DOWNLOADS]: The skill fetches component definitions and documentation from the vendor's official domain (zardui.com) and references commits on the vendor's GitHub repository to verify version history. It also downloads the zard-cli package from the public NPM registry at runtime.
[INDIRECT_PROMPT_INJECTION]: The skill operates by analyzing project files and remote registry data, creating an interface where untrusted data could influence agent behavior.
Ingestion points: Local configuration files (components.json), component source code (.ts, .html), and remote registry files (registry.json) or documentation from zardui.com.
Boundary markers: None explicitly specified in the instructions for delimiting external content during processing.
Capability inventory: Access to shell execution (zard-cli) and file system modification (overwriting components).
Sanitization: No specific sanitization or filtering logic is described for the content ingested from the project or registry.

zard-migration