Tabletopkit logo

Tabletopkit

CommunityPopular
dpearson2699
tabletopkit

Builds multiplayer spatial board games using TabletopKit on visionOS. Use when creating tabletop game experiences with boards, pieces, cards, or dice; managing seats, turns, equipment state, TabletopAction flows, or TabletopInteraction delegates; synchronizing gameplay through FaceTime Group Activities; rendering with RealityKit; or implementing snapping, tosses, and physics on a virtual table surface.

Overview

Publisherdpearson2699
Repositoryswift-ios-skills
Skill nametabletopkit
Stars
1.2K
Forks
61
Bundled files
2
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.

  • 2 bundled files

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

  • Open source

    Published by dpearson2699 on GitHub. Read the source before you install it.

Installation

Install the Tabletopkit 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/dpearson2699/swift-ios-skills/tree/main/skills/tabletopkit
  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/dpearson2699/swift-ios-skills.git /tmp/swift-ios-skills
mkdir -p .claude/skills
cp -r /tmp/swift-ios-skills/skills/tabletopkit .claude/skills/tabletopkit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Tabletopkit 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 Tabletopkit 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 Tabletopkit 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.

TabletopKit

Build visionOS board games whose synchronized state changes flow through TabletopAction and render with RealityKit. The availability matrix below owns version details.

Contents

Setup

TierAPIs
visionOS 2.0+Core gameplay, equipment, seats, actions, rendering, Group Activities
visionOS 2.2+TabletopInteraction.Configuration
visionOS 26.0+Custom actions/state, registration, advanced toss outcomes, discarded-action observation

Simulator supports single-player layout testing, not multiplayer.

Project Configuration

  1. import TabletopKit in source files that define game logic.
  2. import RealityKit for entity-based rendering.
  3. For multiplayer, add the Group Activities capability in Signing & Capabilities.
  4. Provide table, piece, card, and dice USDZ assets in a RealityKit content bundle.

Key Types Overview

TypeRole
TabletopGameCentral game manager; owns setup, actions, observers, rendering
TableSetupConfiguration object passed to TabletopGame init
Tabletop / EntityTabletopProtocol for the table surface
Equipment / EntityEquipmentProtocol for interactive game pieces
TableSeat / EntityTableSeatProtocol for player seat positions
TabletopActionCommands that modify game state
TabletopInteractionGesture-driven player interactions with equipment
TabletopGame.ObserverCallback protocol for reacting to confirmed actions
TabletopGame.RenderDelegateCallback protocol for visual updates
EntityRenderDelegateRealityKit-specific render delegate

Game Configuration

Build and validate a game in this order:

  1. Define the tabletop, equipment, and seats.
  2. Configure TableSetup and register every custom action type.
  3. Create the game, attach its observer and renderer, claim a seat, and establish automatic or manual update handling.
  4. Inspect the current snapshot for required equipment IDs, parents, seats, and counters before starting multiplayer. Fix the setup and rebuild if an invariant fails.
swift
import TabletopKit
import RealityKit

let table = GameTable()
var setup = TableSetup(tabletop: table)
setup.add(seat: PlayerSeat(index: 0, pose: seatPose0))
setup.add(seat: PlayerSeat(index: 1, pose: seatPose1))
setup.add(equipment: GamePawn(id: .init(1)))
setup.add(equipment: GameDie(id: .init(2)))

let game = TabletopGame(tableSetup: setup)
game.claimAnySeat()

Call update(deltaTime:) each frame if automatic updates are not enabled via the .tabletopGame(_:parent:automaticUpdate:) modifier. Read state safely with withCurrentSnapshot(_:).

Table and Board

Tabletop Protocol

Conform to EntityTabletop to define the playing surface. Provide a shape (round or rectangular) and a RealityKit Entity for visual representation.

swift
struct GameTable: EntityTabletop {
    var shape: TabletopShape
    var entity: Entity
    var id: EquipmentIdentifier

    init() {
        entity = try! Entity.load(named: "table/game_table", in: contentBundle)
        shape = .round(entity: entity)
        id = .init(0)
    }
}

Table Shapes

Use factory methods on TabletopShape:

swift
// Round table from dimensions
let round = TabletopShape.round(
    center: .init(x: 0, y: 0, z: 0),
    radius: 0.5,
    thickness: 0.05,
    in: .meters
)

// Rectangular table from entity
let rect = TabletopShape.rectangular(entity: tableEntity)

Equipment (Pieces, Cards, Dice)

Equipment Protocol

All interactive game objects conform to Equipment (or EntityEquipment for RealityKit-rendered pieces). Each piece has an id (EquipmentIdentifier) and an initialState property.

Choose the state type based on the equipment:

State TypeUse Case
BaseEquipmentStateGeneric pieces, pawns, tokens
CardStatePlaying cards (tracks faceUp / face-down)
DieStateDice with an integer value
RawValueStateCustom data encoded as UInt64
CustomEquipmentStateCustom state with a BaseEquipmentState plus game data; see the availability matrix

Defining Equipment

swift
// Pawn -- uses BaseEquipmentState
struct GamePawn: EntityEquipment {
    var id: EquipmentIdentifier
    var initialState: BaseEquipmentState
    var entity: Entity

    init(id: EquipmentIdentifier) {
        self.id = id
        self.entity = try! Entity.load(named: "pieces/pawn", in: contentBundle)
        self.initialState = BaseEquipmentState(
            parentID: .init(0), seatControl: .any,
            pose: .identity, entity: entity
        )
    }
}

// Card -- uses CardState (tracks faceUp)
struct PlayingCard: EntityEquipment {
    var id: EquipmentIdentifier
    var initialState: CardState
    var entity: Entity

    init(id: EquipmentIdentifier) {
        self.id = id
        self.entity = try! Entity.load(named: "cards/card", in: contentBundle)
        self.initialState = .faceDown(
            parentID: .init(0), seatControl: .any,
            pose: .identity, entity: entity
        )
    }
}

// Die -- uses DieState (tracks integer value)
struct GameDie: EntityEquipment {
    var id: EquipmentIdentifier
    var initialState: DieState
    var entity: Entity

    init(id: EquipmentIdentifier) {
        self.id = id
        self.entity = try! Entity.load(named: "dice/d6", in: contentBundle)
        self.initialState = DieState(
            value: 1, parentID: .init(0), seatControl: .any,
            pose: .identity, entity: entity
        )
    }
}

ControllingSeats

Restrict which players can interact with a piece via seatControl:

  • .any -- any player
  • .restricted([seatID1, seatID2]) -- specific seats only
  • .restrictedCurrent([seatID1, seatID2]) -- specific seats only while they are in turn
  • .current -- only the seat whose turn it is
  • .inherited -- inherits from parent equipment

Equipment Hierarchy and Layout

Equipment can be parented to other equipment. Override layoutChildren(for:visualState:) to position children. Return one of:

  • .planarStacked(layout:animationDuration:) -- cards/tiles stacked vertically
  • .planarOverlapping(layout:animationDuration:) -- cards fanned or overlapping
  • .volumetric(layout:animationDuration:) -- full 3D layout

See references/tabletopkit-patterns.md for card fan, grid, and overlap layout examples.

Player Seats

Conform to EntityTableSeat and provide a pose around the table:

swift
struct PlayerSeat: EntityTableSeat {
    var id: TableSeatIdentifier
    var initialState: TableSeatState
    var entity: Entity

    init(index: Int, pose: TableVisualState.Pose2D) {
        self.id = TableSeatIdentifier(index)
        self.entity = Entity()
        self.initialState = TableSeatState(pose: pose, context: 0)
    }
}

Claim a seat before interacting: game.claimAnySeat(), game.claimSeat(matching:), or game.releaseSeat(). Observe changes via TabletopGame.Observer.playerChangedSeats.

Game Actions and Turns

Built-in Actions

Use TabletopAction factory methods to modify game state:

swift
// Move equipment to a new parent
game.addAction(.moveEquipment(matching: pieceID, childOf: targetID, pose: newPose))

// Flip a card face-up
game.addAction(.updateEquipment(card, faceUp: true))

// Update die value
game.addAction(.updateEquipment(die, value: 6))

// Set whose turn it is
game.addAction(.setTurn(matching: TableSeatIdentifier(1)))

// Update a score counter
game.addAction(.updateCounter(matching: counterID, value: 100))

// Create a state bookmark (for undo/reset)
game.addAction(.createBookmark(id: StateBookmarkIdentifier(1)))

Custom Actions

For game-specific logic, conform to CustomAction when the availability matrix permits it. Custom action application and validation must depend only on the action data and the supplied TableState / TableSnapshot so every peer resolves the same result. Register custom action types during setup before dispatching them:

swift
setup.register(action: CollectCoin.self)
game.addAction(CollectCoin(coinID: coinID, playerID: playerID))

Register each custom action type before dispatching it. See references/tabletopkit-patterns.md for full custom action and custom state examples.

Score Counters

swift
setup.add(counter: ScoreCounter(id: .init(0), value: 0))
// Update: game.addAction(.updateCounter(matching: .init(0), value: 42))
// Read:   snapshot.counter(matching: .init(0))?.value

State Bookmarks

Save and restore game state for undo/reset:

swift
game.addAction(.createBookmark(id: StateBookmarkIdentifier(1)))
game.jumpToBookmark(matching: StateBookmarkIdentifier(1))

Bookmark restoration is asynchronous and network ordered. Wait for stateDidResetToBookmark, then read withCurrentSnapshot, rebuild local UI from that authoritative callback state, and validate the restored invariant. Do not inspect immediately after jumpToBookmark or enqueue another jump on mismatch. Load State Bookmarks and Undo for the observer reconciliation pattern.

Interactions

TabletopInteraction.Delegate

Return an interaction delegate from the .tabletopGame modifier to handle player gestures on equipment:

swift
.tabletopGame(game.tabletopGame, parent: game.renderer.root) { value in
    if game.tabletopGame.equipment(of: GameDie.self, matching: value.startingEquipmentID) != nil {
        return DieInteraction(game: game)
    }
    return DefaultInteraction(game: game)
}

Use interaction.value.gesture for gesture-specific state. Avoid deprecated gesturePhase. For destination control, prefer interaction.setConfiguration(.init(allowedDestinations: ...)) when available rather than deprecated setAllowedDestinations(_:) or value.allowedDestinations.

Handling Gestures and Tossing Dice

Basic toss(equipmentID:as:) is core TabletopKit; the availability matrix owns advanced toss outcomes.

swift
class DieInteraction: TabletopInteraction.Delegate {
    let game: Game

    func update(interaction: TabletopInteraction) {
        switch interaction.value.phase {
        case .started:
            interaction.setConfiguration(.init(allowedDestinations: .any))
        case .update:
            if interaction.value.gesture?.phase == .ended {
                interaction.toss(
                    equipmentID: interaction.value.controlledEquipmentID,
                    as: .cube(height: 0.02, in: .meters)
                )
            }
        case .ended, .cancelled:
            break
        }
    }

    func onTossStart(interaction: TabletopInteraction,
                     outcomes: [TabletopInteraction.TossOutcome]) {
        for outcome in outcomes {
            let face = outcome.tossableRepresentation.face(for: outcome.restingOrientation)
            interaction.addAction(.updateEquipment(
                die, rawValue: face.rawValue, pose: outcome.pose
            ))
        }
    }
}

Tossable Representations

Dice physics shapes: .cube (d6), .tetrahedron (d4), .octahedron (d8), .decahedron (d10), .dodecahedron (d12), .icosahedron (d20), .sphere. All take height:in: (or radius:in: for sphere) and optional restitution:.

Programmatic Interactions

Start interactions from code: game.startInteraction(onEquipmentID: pieceID).

See references/tabletopkit-patterns.md for group toss, predetermined outcomes, interaction acceptance/rejection, and destination restriction patterns.

RealityKit Rendering

Conform to EntityRenderDelegate to bridge state to RealityKit. Provide a root entity. TabletopKit automatically positions EntityEquipment entities.

swift
class GameRenderer: EntityRenderDelegate {
    let root = Entity()

    func onUpdate(timeInterval: Double, snapshot: TableSnapshot,
                  visualState: TableVisualState) {
        // Custom visual updates beyond automatic positioning
    }
}

Connect to SwiftUI with .tabletopGame(_:parent:automaticUpdate:) on a RealityView:

swift
struct GameView: View {
    let game: Game

    var body: some View {
        RealityView { content in
            content.entities.append(game.renderer.root)
        }
        .tabletopGame(game.tabletopGame, parent: game.renderer.root) { value in
            GameInteraction(game: game)
        }
    }
}

Debug outlines: game.tabletopGame.debugDraw(options: [.drawTable, .drawSeats, .drawEquipment])

Group Activities Integration

TabletopKit integrates directly with GroupActivities for FaceTime-based multiplayer. Define a GroupActivity, then call coordinateWithSession(_:). TabletopKit automatically synchronizes all equipment state, seat assignments, actions, and interactions. No manual message passing required.

swift
import GroupActivities

struct BoardGameActivity: GroupActivity {
    var metadata: GroupActivityMetadata {
        var meta = GroupActivityMetadata()
        meta.type = .generic
        meta.title = "Board Game"
        return meta
    }
}

@Observable
class GroupActivityManager {
    let tabletopGame: TabletopGame
    private var sessionTask: Task<Void, Never>?

    init(tabletopGame: TabletopGame) {
        self.tabletopGame = tabletopGame
        sessionTask = Task { @MainActor in
            for await session in BoardGameActivity.sessions() {
                tabletopGame.coordinateWithSession(session)
            }
        }
    }

    deinit { tabletopGame.detachNetworkCoordinator() }
}

Implement TabletopGame.MultiplayerDelegate for joinAccepted(), playerJoined(_:), didRejectPlayer(_:reason:), and multiplayerSessionFailed(reason:). See references/tabletopkit-patterns.md for custom network coordinators and arbiter role management.

Common Mistakes

  • Skipping seat claim. Players must call claimAnySeat() or claimSeat(_:) before interacting with equipment. Without a seat, actions are rejected.
  • Mutating state outside actions. All state changes must go through TabletopAction or CustomAction. Directly modifying equipment properties bypasses synchronization.
  • Missing custom action registration. Register every custom action with setup.register(action:) before use.
  • Not handling action rollback. Actions are optimistically applied and can be rolled back if validation fails on the arbiter. Implement actionWasRolledBack(_:snapshot:) to revert UI state.
  • Ignoring discarded actions when available. Implement actionWasDiscarded(_:) when local action queue pressure matters; it is called for local actions that cannot be enqueued.
  • Using wrong parent ID. Equipment parentID in state must reference a valid equipment ID (typically the table or a container). An invalid parent causes the piece to disappear.
  • Ignoring TossOutcome faces. After a toss, read the face from outcome.tossableRepresentation.face(for: outcome.restingOrientation) rather than generating a random value. The physics simulation determines the result.
  • Testing multiplayer in Simulator. Group Activities do not work in Simulator. Multiplayer requires physical Apple Vision Pro devices on a FaceTime call.

Review Checklist

  • The centralized platform/availability matrix is applied
  • TableSetup created with a Tabletop/EntityTabletop conforming type
  • All equipment conforms to Equipment or EntityEquipment with correct state type
  • Seats added and claimAnySeat() / claimSeat(_:) called at game start
  • All custom actions registered with setup.register(action:)
  • TabletopGame.Observer reconciles confirmed, rolled-back, discarded, and bookmark-reset outcomes with the current snapshot
  • EntityRenderDelegate or RenderDelegate connected
  • .tabletopGame(_:parent:automaticUpdate:) modifier on RealityView
  • GroupActivity defined and coordinateWithSession(_:) called; multiplayer described as Group Activities/SharePlay synchronization
  • Group Activities capability added in Xcode for multiplayer builds
  • Debug visualization (debugDraw) disabled before release
  • Device notes state Simulator is single-player only; multiplayer requires 2+ Apple Vision Pro units on FaceTime

References

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 Tabletopkit AI skill do?

Builds multiplayer spatial board games using TabletopKit on visionOS. Use when creating tabletop game experiences with boards, pieces, cards, or dice; managing seats, turns, equipment state, TabletopAction flows, or TabletopInteraction delegates; synchronizing gameplay through FaceTime Group Activities; rendering with RealityKit; or implementing snapping, tosses, and physics on a virtual table surface.

Why use Tabletopkit on TypingMind?

Because you install it once and use it with any model. Tabletopkit 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 Tabletopkit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/dpearson2699/swift-ios-skills/tree/main/skills/tabletopkit. 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 Tabletopkit?

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 Tabletopkit?

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

Is the Tabletopkit AI skill free?

It is published on GitHub by dpearson2699. 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 👇