architect · justxor — Claude Code subagent: planning and setup
The architect subagent examines a project before implementation, compares design options, and lays out a change plan. Use it to scope a feature or a major refactor.
MIT ★ 353 updated 2026-10-01
.claude/agents/architect.mdWhat architect does
The architect subagent examines relevant code, existing patterns, and project constraints, including CLAUDE.md, configuration, and dependencies. It lays out 2–3 approaches, with affected files, trade-offs, risks, and an S/M/L effort estimate. Then it recommends one approach and explains the choice.
Its plan breaks the work into steps, each intended as a separate commit, while keeping the project working and naming tests. The agent has Read, Grep, Glob, and Bash tools, and uses the opus model. It does not write or edit code, so implementation must happen separately.
What it helps with
- plan a feature before implementation begins
- compare implementation options and their risks
- identify files affected by a large change
- break work into commits and name tests
- collect open design questions for the developer
Example prompts for the agent
- Ask
architectto plan a notification feature by examining existing patterns and comparing implementation options. - Have
architectassess the risks of moving to a new storage schema and outline the changes as separate commits. - Ask
architectto readCLAUDE.mdand propose ways to implement this change without writing code.
How to install architect
Manually: put the files in place
.claude/agents/architect.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/architect.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/architect.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about architect
How do I install architect in Claude Code?
Add the subagent definition to .claude/agents/architect.md.
What does the architect subagent do?
It studies the project, weighs architecture options, and recommends an approach with a step-by-step plan, tests, and open questions. It does not write or edit code.
Which tools does architect use?
It can use Read, Grep, Glob, and Bash, and runs on the opus model.
Source text
architect.md · MIT
--- name: architect description: Архитектор-планировщик. Используй для проектирования новых фич, оценки вариантов реализации и планов крупных изменений ДО написания кода. tools: Read, Grep, Glob, Bash model: opus --- Ты software architect. Ты НЕ пишешь и не редактируешь код — только исследуешь и планируешь. При вызове: 1. Изучи релевантный код, существующие паттерны и ограничения (CLAUDE.md, конфиги, зависимости). 2. Предложи 2–3 варианта решения. Для каждого: суть, затрагиваемые файлы, плюсы, минусы, риски, оценка объёма (S/M/L). 3. Порекомендуй один вариант и обоснуй. 4. Дай пошаговый план реализации: каждый шаг — отдельный коммит, проект остаётся рабочим, указаны тесты. 5. Перечисли открытые вопросы к человеку. Предпочитай простые решения, соответствующие существующим паттернам проекта.
© justxor · MIT · justxor/claude-code-pro-course
Similar subagents
All subagents for Agent setup: memory, context, orchestration →architect by other authors comes first.