# Buildcrew Think

> Product thinking with the buildcrew thinker — 6 forcing questions that test whether something is worth building, ending in a design doc. Use when the user has a product idea, asks "is this worth building", or wants to think a concept through before code.

- Skill: `z1nun/buildcrew-think` (Agent Skill)
- Install (CLI): `npx skillmds@latest add z1nun/buildcrew-think`
- Raw SKILL.md: https://api.skillmd.com/api/skills/z1nun/buildcrew-think/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: z1nun (https://skillmd.com/u/z1nun)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/z1nun/buildcrew-think

---


# buildcrew-think — Product Thinking

Full role prompt: `.claude/agents/thinker.md`. Read it before executing.
With subagent support, dispatch as a subagent; without, adopt the role in-context.

## Pipeline

```
thinker (single run, interactive with the user)
```

This mode is a conversation, not a monologue — ask the forcing questions one at a
time and build on the user's answers.

## Rules

- Read `.claude/harness/project.md` and `glossary.md` first if present.
- Do not write implementation code in this mode. The deliverable is a design doc.
- If the conclusion is "worth building", offer to continue into the feature
  pipeline (`buildcrew` skill / `@buildcrew`).
- End with: verdict (build / don't build / reshape), the design doc, and open
  questions that remain.

