agents-architect-skill

name	agents-architect:skill
description	Author a single skill from a short brief. Use to add a skill to an existing plugin or iterate on a skill's description/body.
argument-hint	<skill-name> "<one-line purpose>" [--plugin <path>] [--siblings <a,b,c>]
allowed-tools	["Read","Write","Task","AskUserQuestion"]

Fast path to a single, well-triggered skill without running full /agents-architect:new.

<execution_context> @${CLAUDE_PLUGIN_ROOT}/agents-architect/templates/skill.template.md </execution_context>

Arguments: $ARGUMENTS — name, purpose, optional plugin path and sibling skills for disambiguation. 1. Parse args. If plugin path omitted, default to current working directory. 2. `AskUserQuestion` to fill gaps: target trigger phrases (10-30), examples of prompts that should fire it, examples that should NOT. 3. Spawn `agents-architect-skill-author` with the filled brief. It writes `skills//SKILL.md`, `reference.md` (if needed), `evals/`. 4. Report back the produced files + trigger surface. Paths of produced files + 5-line summary.