Skip to content

Think: Design Before You Build

Turns rough ideas into approved, decision-complete plans before any code is written.

Repository
tw93/Waza
Stars
7,493
Licence
MIT
Files
4 files

What it does

When installed, the agent stops jumping straight into code when you describe a feature or problem. Instead it asks for approval on a plan first: the goal, constraints, chosen approach, rejected alternatives, tests, and handoff steps. It checks existing project docs, config values and official solutions before recommending a design, and it gives a clear opinion rather than a list of vague options. It can also switch to an evaluation mode for keep-or-remove decisions and a triage mode for judging a batch of requests.

Great for

  • Developers planning a new feature or architecture before writing code
  • Teams deciding whether an idea is worth building
  • People who want a written handoff plan an agent or teammate can execute without re-deciding

Skip it if

  • Simple bug fixes or small edits, which the skill itself says it is not for, since the extra planning step adds delay
  • Non-technical users who only want finished output such as a document or design, since the skill is about planning, not producing deliverables

Try asking

  • “Plan this: add offline sync to our notes app, what approach should we take?”
  • “Should we keep the legacy export feature, or is it worth removing?”
  • “How should I design a rate limiter for our public API?”
  • “Here are five feature requests from users, which are worth doing?”
Step-by-step guide: how to use Think: Design Before You BuildInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 4 original files, the MIT licence and an attribution note.

    think.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip think.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip think.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent the real repo context and point it at your AGENTS.md or CLAUDE.md so it can check project rules before planning.
  • Answer the approval question explicitly; the skill is designed to wait for your sign-off before any implementation.
  • Ask for a lightweight plan when the problem is already defined, and a full plan when you have several competing approaches.

What's inside

  • references/durable-context.md
  • references/mode-evaluation.md
  • references/mode-triage.md
  • SKILL.md

Redistributed unmodified from tw93/Waza under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.