hsb3/agent-interface
Design and check CLI command and MCP tool names against a noun-verb standard, and how a CLI behaves for agents: output, errors, paging, prompts and safe retries.
Check and design how a CLI behaves when an agent drives it: machine output (--json and one envelope), what goes to stdout and stderr, error shape and codes, exit status, paging, prompts and confirmation, dry run and safe retries. Use when the user asks to check how a CLI behaves for agents, asks whether a CLI is safe or ready for agents to drive, asks to check their CLI without saying names or behavior ("check my CLI", "is my CLI ready for agents"), asks to accept or decline an option from a behavior plan, or asks to design how a new CLI should behave for agents. Do not use for command or tool names (that is the menu-design skill), for ordinary CLI use such as running a command or showing git status, for building an MCP server, or for unrelated questions.
Design and check CLI command names and MCP tool names against a noun-verb naming standard, and plan renames that list every caller. Use when the user asks to check a CLI's command names, asks whether a CLI is noun-verb, asks to design the commands for a new CLI, asks to plan renames for or fix the verbs in a CLI, or asks to name or check MCP tools. Do not use for ordinary CLI use such as listing files or running a command, for building an MCP server or a multi-entry-point service (the agent-toolkits plugin's job), for writing skills, or for unrelated questions. Not for how a CLI behaves for agents (output, errors, paging, prompts); that is the cli-behavior skill.