Gat Design logo

Gat Design

Community
Yuki001
gat-design

Design one system or all remaining systems within a specific planned GAT milestone, producing milestone-scoped GDD, optional content data, and art documents while updating progress and auditing consistency. Requires an explicit milestone.

Overview

PublisherYuki001
Repositorygame-dev-skills
Skill namegat-design
Stars
104
Forks
13
Bundled files
1
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Yuki001 on GitHub. Read the source before you install it.

Installation

Install the Gat Design AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Yuki001/game-dev-skills/tree/main/gat/.agents/skills/gat-design
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Yuki001/game-dev-skills.git /tmp/game-dev-skills
mkdir -p .claude/skills
cp -r /tmp/game-dev-skills/gat/.agents/skills/gat-design .claude/skills/gat-design
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Gat Design in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Gat Design on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Gat Design is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Design

Codex Runtime

This skill is fully contained in the installed .agents directory. Resolve all linked resources relative to this SKILL.md; do not read workflow resources outside this packaged directory.

Before acting, read GAT Workflow. Read Directory Structure or Workflow Catalog only when their structure or ownership details are needed.

Use these packaged role profiles for every role-specific pass:

When the workflow says to spawn one of these roles, create a Codex subagent with that exact task name when collaboration is available. Give it the complete role profile, relevant source documents, confirmed decisions, template, and exact target paths; wait for it and review its result before continuing. This skill explicitly authorizes those named workflow subagents. If subagents are unavailable, perform the same pass yourself using the packaged role profile.

Treat the directory containing .agents as the repository root. All GAT project documents live under that root's gat/ directory. Use Codex filesystem tools and the environment's supported patch editor while preserving unrelated user changes.

Interpret AskUserQuestion in the procedure as a single concise user question. Use a structured choice UI only for a genuine 2-3 option decision; otherwise ask in ordinary dialogue and wait. Command examples use Codex skill syntax ($gat-*).

This skill designs a single system within a specific milestone, after $gat-brainstorm (overview) and $gat-milestone (milestone plan) are done. For story, worldbuilding, characters, quests, dialogue, or authored narrative content, delegate to $gat-story. For the initial concept interview, delegate to $gat-brainstorm.

The milestone argument is mandatory. If it is omitted, the skill stops and asks the user to specify a milestone — no inference, default, or fallback.

Phase 1: Resolve Mode

Check existing files:

  • If gat/overview/game.md or gat/overview/systems-index.md missing → hand off: tell the user to run $gat-brainstorm [hint] first. Stop.
  • If gat/overview/art-direction.md missing → hand off: tell the user to run $gat-brainstorm [hint] to establish global art direction. Stop.
  • If gat/milestone/milestone.md missing → hand off: tell the user to run $gat-milestone first. Stop.
  • If no milestone argument provided → STOP. Tell the user a milestone is required and show the milestones listed in gat/milestone/milestone.md. Do NOT infer or default one. Do NOT write any files.
  • If the milestone argument does not match any milestone directory under gat/milestone/ (i.e., no m{N}-<name>/m{N}-brief.md exists) → STOP. Tell the user the milestone was not found and list the valid milestones.

Then resolve sub-mode:

  • If a system (or hint) argument is also provided → Mode: system (design that one system in the milestone)
  • If only a milestone argument is provided → Mode: continue (auto-fill all systems in that milestone's brief that are not yet designed). Continue is scoped to that milestone only — it does NOT cross into other milestones.

Phase 2: Execute

Read templates:

  • ../../assets/templates/design/system-gdd.md
  • ../../assets/templates/design/system-art.md
  • ../../assets/templates/design/content-data.md

Read context:

  • gat/overview/game.md, gat/overview/systems-index.md, gat/overview/art-direction.md
  • the target milestone's m{N}-brief.md (for scope, in-scope systems, and the progress tracker)
  • relevant gat/narrative/*.md when narrative context affects the system
  • existing system docs under gat/milestone/m{N}-<name>/<system>/ if present

Also read existing gat/narrative/*.md when present and pass relevant narrative context to spawned agents. Do not create or rewrite narrative docs here; use $gat-story for that.

Continue mode ($gat-design {milestone})

For each system listed in the milestone's m{N}-brief.md that is not yet designed (progress row Pending), in the brief's order:

  1. Spawn gat-designer → gat/milestone/m{N}-<name>/<system>/<system>-gdd-m{N}.md
  2. If the system needs content data, spawn gat-designer → gat/milestone/m{N}-<name>/<system>/<system>-data-m{N}.md
  3. Spawn gat-artist → gat/milestone/m{N}-<name>/<system>/<system>-art-m{N}.md
  4. Update that system's progress row in m{N}-brief.md to Designed.

Do NOT design systems belonging to other milestones.

System mode ($gat-design {milestone} {system | hint})

Step 1 — Spawn gat-designer → gat/milestone/m{N}-<name>/<system>/<system>-gdd-m{N}.md

  • Pass gat/overview/game.md, gat/overview/systems-index.md, the milestone's m{N}-brief.md, system-gdd template
  • Pass existing system GDD if present (e.g., an earlier milestone's version of the same system, for reference — a later milestone may redefine it differently)

Step 2 — Content Fill (for systems with high content volume, within this milestone's scope)

If the system requires substantial content data, spawn gat-designer to write gat/milestone/m{N}-<name>/<system>/<system>-data-m{N}.md:

  • Pass the system GDD, gat/overview/game.md, gat/overview/systems-index.md, the milestone brief, content-data template
  • Pass existing content doc if present
  • The content doc fills specific instances, parameters, sequences, and groups for THIS milestone — the data that instantiates the rules defined in the milestone's system GDD

A system needs a content-data doc when its GDD defines data structures that need many concrete instances (e.g. an enemy GDD defines enemy attributes → content doc fills the specific enemies in scope for this milestone).

Systems that are purely mechanical (e.g. input, tbs-scoring) typically do NOT need content-data docs — their parameters fit within the GDD itself.

Step 3 — Spawn gat-artist → gat/milestone/m{N}-<name>/<system>/<system>-art-m{N}.md:

  • Pass gat/overview/game.md, gat/overview/art-direction.md, the milestone's system GDD, the content-data doc (if it exists), the milestone brief, system-art template
  • Pass existing system art doc if present

Step 4 — Update Progress

After writing a system's docs, update that system's row in the milestone's m{N}-brief.md progress tracker (System | Iteration(s) | GDD | Art | Data | Design Status) to mark GDD/Art/Data paths filled and Design Status Designed. If this was the first system designed in the milestone, also set the milestone status from planned to designing (in both the brief and gat/milestone/milestone.md).

Phase 3: Cross-Document Consistency Audit

After ALL documents in the current request are written (all systems in continue mode, or the single system in system mode), run a cross-document consistency audit before the review summary.

Scope — audit every newly written or updated document in this request against:

  • Every other newly written document in this request (peer conflict check)
  • Existing system GDD/docs within the SAME milestone that were NOT touched in this request (regression check — does the new design contradict an already- designed system?)
  • gat/overview/game.md and gat/overview/systems-index.md (global coherence)
  • gat/overview/art-direction.md (art coherence, for system-art docs)

Checklist — for each document pair, verify:

  1. Terminology: same concept, same name everywhere. No synonyms or overlapping terms across systems.
  2. Scope boundaries: no two systems claim ownership of the same responsibility, data, or behavior.
  3. Data flow: referenced data/events/signals that cross system boundaries are defined on the producing side and match the consuming side's expectation.
  4. Parameter consistency: numeric parameters (speeds, timings, counts, thresholds) that appear in multiple docs are identical or intentionally different with a documented reason.
  5. Dependency order: if system A depends on system B, system B's design must exist (or be designed in this same request) and must provide what A consumes.
  6. Narrative alignment: if gat/narrative/*.md exists, check that system behaviors don't contradict established story/world/character facts.

Resolution — for each conflict found:

  • If the conflict is unambiguous (clear which side is correct), fix the offending document(s) directly and note the fix in the audit summary.
  • If the conflict requires a design decision, use AskUserQuestion to present both sides and ask the user to decide. Wait for the user's answer, then resolve the conflict accordingly and update the affected document(s).

Output — add an "Audit" section at the bottom of the review summary (Phase 4) listing every check performed (document pairs audited), every conflict found, and how each was resolved (fixed / flagged for user).

Phase 4: Review

Summarize what was created or updated for the system(s) in the milestone.

Use AskUserQuestion:

  • Design another system in this milestone → Run $gat-design <milestone> <system>
  • Continue remaining systems in this milestone → Run $gat-design <milestone>
  • Stop here (hand the milestone to engineering when all its systems are Designed)

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Gat Design AI skill do?

Design one system or all remaining systems within a specific planned GAT milestone, producing milestone-scoped GDD, optional content data, and art documents while updating progress and auditing consistency. Requires an explicit milestone.

Why use Gat Design on TypingMind?

Because you install it once and use it with any model. Gat Design is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Gat Design in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Yuki001/game-dev-skills/tree/main/gat/.agents/skills/gat-design. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Gat Design?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Gat Design?

As many as you like. As long as a model supports skills, you can use Gat Design with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Gat Design AI skill free?

It is published on GitHub by Yuki001. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇