CLI/Agent/ic/Principle

• Agentic CLI Principles

  • Project Overview

    • Thesis

      • Engineers shouldn't design each new CLI/Tool to be a unique snowflake in terms of its structure, flags, documentation, etc. Instead, engineers and AI/Agents should maintain a standard for creating CLIs that measurably improves both human and AI performance at using that CLI.

    • Goal

      • IF

        • A Developer is making a new CLI tool and wants to optimize for the DX and AI/Agent/AX AI agents that will use that

      • WHEN

        • the developer points their AI/Coding/Tool at the project's repo

      • THEN

        • the coding agent will

          • design a CLI that follows the standard

            • which is specifically designed to measurably help agents as well as humans use that CLI
            • familiarity with one CLI that follows the standard should help agents and humans use all CLIs that follow the standard

          • measure the designed CLI against its standard

            • evaluate the adherence to and the performance of agent usage of the CLI

  • Project Elaboration

    • A git repository should exist that helps foster best practices / principles of the CLIs.
    • It should contain a Knowledge Garden maintained by AI Agents that research and evaluate precedents for agentic CLIs.
      • The precedents should include
        • 1 - case studies of existing CLIs
        • 2 - other Standards related to CLIs