# Design Research

> Search for how other games solve a specific design problem — game loops, mechanics, systems, progression, economy. Returns a comparison of 2-4 reference games with what to borrow. Use when designing a new feature, exploring game mechanics, or looking for design inspiration before writing a design doc.

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

---


Respond in the user's preferred language (detect from their recent messages, or fall back to the language setting in CLAUDE.md).

Search for **game design** references to the problem currently being discussed. Focus on design patterns and player experience, not code.

Arguments: $ARGUMENTS

## Process

1. Identify the core design question (e.g., "how to structure a survival loop on a ship", "how trading between cities creates interesting choices")
2. Search for game design analyses, GDC talks, postmortems, and design breakdowns using WebSearch. Good query patterns:
   - `"[game name]" game design analysis [mechanic]`
   - `GDC [mechanic] design postmortem`
   - `[mechanic] game design breakdown`
3. Organize into a comparison table:

| Game | How they do it | What makes it fun | What we could borrow |
|---|---|---|---|

4. Recommend which approach best fits this project's scope and feel, and why

## Notes

- Focus on the **player experience**, not implementation — "RimWorld uses a bill queue per workbench that lets players set min/max thresholds" is good; "WorkGiver_DoBill class" is not
- Prefer games the team has likely played or that are well-documented
- Include at least one game from a different genre if it solves the same design problem in an interesting way
- If searching yields thin results, lean on your own knowledge of game design — but flag which insights came from search vs. your own analysis