2-step ๋์์ธ ํ์ดํ๋ผ์ธ: --step brief (ํ๊ฒ ๋ถ์ โ DESIGN_BRIEF.md ์์ฑ) โ --step token (ํ ํฐ/DESIGN.md ์์ฑ). --step all์ด ๊ธฐ๋ณธ๊ฐ.
Design an agent memory system โ working memory, episodic memory, semantic memory, and procedural memory. Use when building agents that need to remember context across sessions, learn from interactions, or maintain persistent knowledge. Covers storage strategy, retrieval patterns, and context management.
Select and design the right orchestration pattern for multi-agent systems. Evaluate Sequential, Parallel, Router, and Hierarchical patterns against your use case requirements. The Router pattern covers both agent selection (classify input to the right specialist) and model selection (route tasks to the right LLM by complexity to balance cost, latency, quality, and fallback chains). Use when deciding how multiple agents should coordinate, share context, delegate tasks, or which model each task should run on.
์ ๋ต ์ค๊ณ ํตํฉ โ ๋น์ฆ๋์ค ๋ชจ๋ธ ์บ๋ฒ์ค(biz-model), ๊ฒฝ์ ํด์ ๋ถ์(moat), ์ฑ์ฅ ๋ฃจํ ์ค๊ณ(growth-loop) ํตํฉ. Use when defining business strategy, competitive positioning, or growth mechanics for an AI product.
์์ด์ ํธ ํ๊ฒฝ ์ค์ ํตํฉ โ 7์์ ์ธ์คํธ๋ญ์ ์์ฑ(agent-instructions)๊ณผ CLAUDE.md/AGENTS.md ๊ตฌ์ฑ(claude-md) ํตํฉ. ์์ด์ ํธ ์ ์ฒด์ฑยท๋๊ตฌยท์ ์ฝยท์คํจ ๋ชจ๋ ์ ์๋ถํฐ ํ๋ก์ ํธ ๋ฉ๋ชจ๋ฆฌ ํ์ผ๊น์ง ํ ์คํฌ๋ก. Use when setting up a new agent or updating agent instructions.
PM์ด ์ฌ๋์๊ฒ ์ง๋ฌธํ๊ณ ๋ต์ ๋ชจ์ผ๋ ๋น๋๊ธฐ ์ฑ๋ โ comms MCP(Gmail/Notion/Zoom/Slack)๋ฅผ ๊ฐ์ธ๋ ๋ฒ์ญ๊ธฐ. --mode ask(์ง๋ฌธ ์ด์ ์์ฑ), --mode pull-answers(์ค๋ ๋ยทํ์๋ก์์ ๋ต ์์ง), --mode digest(์์งํ ๋ต ์์ฝ โ decision-log/ticket-bridge ๋ผ์ฐํ ), --mode review(PRD ์คํ ์ดํฌํ๋ ๋ฆฌ๋ทฐ โ ๋ฆฌ๋ทฐ์ด ๋ฐฐ์ ยท์ฝ๋ฉํธ ์์งยทSignoff audit trail), --mode solo(ํ์ ์์ ๋ Claude๊ฐ ์ญํ ๋๋ฆฌ ์๋ฎฌ๋ ์ด์ ), --mode init(ํ ์ธํ ๋ํํ ์จ๋ณด๋ฉ). ๋ฉ์์ง๋ฅผ ์๋ ๋ฐ์กํ์ง ์๋๋ค โ ์ด์/์ฝ๋ฉํธ๊น์ง๋ง. ์๋ก ์ ์ ๋ solo, ๊ธฐ์ PM์ review๋ก ๋ถ๊ธฐ. Use when a PM needs to ask teammates for status/decisions, run a multi-stakeholder PRD review, and collect answers into hplan.
Run a full product build loop in one orchestrated session โ discover โ research โ design โ PRD โ task decomposition โ team-based implementation. Use when the user invokes /build, when an idea needs end-to-end execution from problem to shipped change, or when a feature crosses discover/architect/deliver boundaries that would otherwise require manual hand-offs.
ํ์คํฌ๋ณ fresh subagent ๋์คํจ์น + 2๋จ๊ณ ๊ฒ์ดํธ(specโquality) ๋ฐ๋ณต ์คํ. harness-plan ์น์ธ ํ ๊ตฌํ ๋ฃจํ๋ฅผ ๋๋ฆด ๋ ์ฌ์ฉ. ๊ธฐ์กด parallel-team(ํ conductor์ ํตํฉ)์ ์ญํ ๋ณ๋ ฌ ๋ฐฉ์๊ณผ ๋ฌ๋ฆฌ, conductor๋ ํ์คํฌ ์์ฐจ+๊ฒ์ดํธ๋ค.