Quick Start

In 15 minutes you will go through the core Brewcode workflow: from writing a structured specification to capturing your project’s conventions and generating a project-tailored deep review skill.

Before you begin

Make sure Claude Code and the brewcode plugin are installed. If not, follow the instructions in Installation.

A skill whose name ends in -setup installs a mechanism you use afterwards instead of the skill itself; recurring tools keep bare names. The full ordered list lives in Full Setup.

Every command below is one you type yourself. Nothing in this suite auto-activates, and each -setup step wants its own fresh session — it asks real questions and spawns a lot of agents.

Full workflow

Install brewcodeplugin install
Add the plugin from the marketplace and install it into your Claude Code session:
```
claude plugin marketplace add https://github.com/kochetkov-ma/claude-brewcode
claude plugin install brewcode@claude-brewcode
```
This step runs once per machine. Reload with /reload-plugins if already inside a session.
Set up the task board/brewtools:task-board-setup
Run it once per repo and accept the spec layer when asked:
```
/brewtools:task-board-setup install . “one lane per service”
```
Then write the spec for a task on the board:
```
/task-spec “Implement JWT authorization with role-based access”
```
What happens: task-board-setup analyzes the repo with parallel agents and deploys a file-based Kanban plus a project-tailored /task-spec skill. /task-spec researches the codebase, asks clarifying questions, and fans out to domain agents for system design.
Result: a task on the board with its own spec and design docs under .claude/features/.
Capture project conventions/brewcode:convention
```
/brewcode:convention full
```
What happens: parallel agents analyze each architectural layer, extract etalon classes and recurring patterns, and write them into .claude/convention/.
Result: every later agent has a written reference for how this project does things — no more guessing at style or structure.
Generate a deep review skill/brewcode:superreview-setup
```
/brewcode:superreview-setup install “QUICK depth; gate on the existing build and lint commands”
```
What happens: analyzes your codebase and conventions, then generates a project-tailored quorum review skill that merges code review and standards checking into one reusable skill installed locally at .claude/skills/.
Result: a custom review skill ready to invoke any time you need a deep, project-aware code review.
Create a domain agent team/brewcode:teams-setup — optional
```
/brewcode:teams-setup install core “one agent per bounded context, plus a migrations specialist”
```
Generates 5-20 project-specific agents with self-selection logic and performance tracking. Skip this step if the default agent set is sufficient for your project. To remove a team later, name it: uninstall core, or purge core to drop its archive and trace files too.

What’s next

After the workflow

Your project is set up for iterative development. Here is what you can do next:

Deep review — generate the project’s own quorum review skill, then invoke the generated /superreview whenever you need it:

/brewcode:superreview-setup status

Extract rules — convert knowledge into project-wide coding rules. This skill takes a free-form instruction, so say what you want:

/brewcode:rules "derive rules from .claude/convention/ and the last 200 commits, one file per domain, no duplicates of CLAUDE.md"

Prompt optimization — reduce tokens in prompts and agents:

/brewtools:text-optimize .claude/agents/

All skills

🏛

Quick Start

Before you begin

Full workflow

What’s next

After the workflow

All skills

brewcode

brewtools

brewui

brewdoc

Full Setup

Dynamic teams

Latest Release

View on GitHub