Use BEFORE writing a full GDD — when the game is still a fuzzy idea ("a roguelite with monster taming", "a city-builder on the moon"). Walks through 7 core questions that crystallize concept into something concrete enough to start a GDD. Output is a 1-page brief that becomes section 1 of game-design-document. Use when starting a new project, pivoting an existing one, or stuck deciding between two directions.
How this skill is triggered — by the user, by Claude, or both
Slash command
/gamedev-toolkit-claude:game-design-brainstormThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Most projects start with a fuzzy idea ("an ARPG but with crafting", "Stardew Valley meets Slay the Spire"). The fuzzy idea phase **must end** before code starts, otherwise you build 3 systems that don't combine into a game.
Most projects start with a fuzzy idea ("an ARPG but with crafting", "Stardew Valley meets Slay the Spire"). The fuzzy idea phase must end before code starts, otherwise you build 3 systems that don't combine into a game.
This skill walks through 7 questions in order. Output is a concrete 1-page brief — concept name, USP (Unique Selling Points), target audience, core loop, perspective, scope, biggest risk. That brief becomes Section 1 of game-design-document.
tdd-with-claude instead)Walk through each. Answer in 1-3 sentences. Don't skip to question 7 — early answers constrain later ones.
Format: [Core Verb] [In Setting] [With Twist].
Examples:
If you can't write this in <25 words, the concept isn't crystallized yet. Iterate.
Format: [Demographic] [Existing game preference] [Why they'd pick yours over the alternative].
Examples:
The "why over the alternative" part is what makes it concrete. "Everyone" or "RPG fans" is too vague.
Each USP must be specific (not "fun combat" — what's unique about your combat). Most games have 2-4 USPs that actually distinguish them; more than 4 means you're losing focus.
Sea Nation USP examples:
SimRoyale USP examples:
Test: would a player describing your game to a friend mention these without prompting? If no, the USP is theoretical not real.
The 30-second loop that the player does over and over. Cover both moment-to-moment and session-level.
Format:
Sea Nation:
SimRoyale:
If your moment-to-moment is "watch numbers go up" or your session-level is "play random levels"... that's not a loop, it's a phase. Iterate.
This decision drives 60% of art budget and feel. Pin it now.
Sea Nation: isometric top-down, mouse+keyboard, free camera with rotation. SimRoyale: isometric, mobile touch, locked camera with auto-follow.
If your team is 1 person, scope is solo-realistic. ARPG endgame is 100+ hours of content = years of work. Pin a realistic scope or you ship nothing.
Every project has one design choice or technical bet that, if it fails, kills the project. Identify it early.
Examples:
The risk drives what you prototype first. Build the riskiest thing in pre-production, not in month 8.
# [Game Name] — concept brief
**One-sentence pitch**: [Q1 answer]
## Target audience
[Q2 answer]
## USP (3 bullets)
- [Q3 first USP]
- [Q3 second USP]
- [Q3 third USP]
## Core loop
- **Moment-to-moment**: [Q4 short loop]
- **Session level**: [Q4 long loop]
## Perspective & controls
[Q5 answer]
## Scope
- Team: [size]
- Timeline: [months]
- Target: [Steam/mobile/F2P/paid/etc.]
- Audience target: [size]
- Price reference: [comparable games + their price]
## Biggest risk
[Q7 answer + first prototype that validates/disproves]
## Status
- [ ] Concept locked
- [ ] Risk prototype done
- [ ] Ready to expand into full GDD (use `game-design-document` skill)
Once the 1-page brief feels right (not just looks right — sleep on it, come back, still feel right?), invoke the game-design-document skill to expand into the full design doc. The 7 answers become Section 1 of the GDD; everything else follows.
If you can't lock all 7 questions, don't expand to a GDD yet. A vague GDD with "TBD" sections is worse than a concrete 1-page brief. Iterate the brief.
npx claudepluginhub marcin-elj/gamedev-toolkit-claude --plugin gamedev-toolkit-claudeGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.