Cli Design

What it does

This skill, Cli Design, generates designs for command-line interfaces (CLIs). It focuses on creating CLIs that are primarily intended to be used by AI agents rather than humans directly. All commands return structured JSON data, which the agent can parse and act upon; humans can optionally use tools like jq to interpret this output. The design emphasizes providing clear "next actions" in each response via a next_actions array, guiding the agent's subsequent steps.

When to use it

• When designing CLIs for automated workflows driven by AI agents.
• For systems where structured data and programmatic control are more important than human readability.
• To create interfaces that provide clear guidance on what actions an agent should take next.
• In situations where a consistent, machine-parseable output format is required.

Key capabilities

• JSON Output: All commands return data in JSON format, with no plain text or tables.
• HATEOAS (Hypermedia as the Engine of Application State): Responses include next_actions – command templates that guide agent behavior.
• Contextual Next Actions: The suggested actions are tailored to the current state and result of a command.
• POSIX/docopt Placeholder Syntax: Templates use standard placeholder syntax for required arguments (<placeholder>), optional flags with values ([--flag <value>]), and boolean flags ([--flag]).

Example prompts

• "Generate a CLI design for managing video downloads, including status checks and log viewing."
• "Design a command to send a pipeline event and provide suggestions for checking its run status and viewing logs."
• "Create a CLI that returns the system health in JSON format with suggested next actions."

Tips & gotchas

• The skill is designed for agent-first CLIs; human readability is secondary.
• All output will be in JSON format, regardless of the request.
• Pay attention to the next_actions array as it dictates the flow of automated tasks.