MouseTrainer.Simulation 1.1.3

<p align="center"> <a href="README.ja.md">日本語</a> | <a href="README.zh.md">中文</a> | <a href="README.es.md">Español</a> | <a href="README.fr.md">Français</a> | <a href="README.hi.md">हिन्दी</a> | <a href="README.it.md">Italiano</a> | <a href="README.pt-BR.md">Português (BR)</a> </p>

<p align="center"><img src="https://raw.githubusercontent.com/mcp-tool-shop-org/brand/main/logos/MouseTrainer/readme.png" alt="MouseTrainer logo" width="400"></p>

<p align="center"> <a href="https://github.com/mcp-tool-shop-org/MouseTrainer/actions/workflows/ci.yml"><img src="https://github.com/mcp-tool-shop-org/MouseTrainer/actions/workflows/ci.yml/badge.svg" alt="CI"></a> <img src="https://img.shields.io/badge/.NET-10-512BD4" alt=".NET 10"> <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-blue.svg" alt="License: MIT"></a> <a href="https://mcp-tool-shop-org.github.io/MouseTrainer/"><img src="https://img.shields.io/badge/Landing_Page-live-blue" alt="Landing Page"></a> </p>

Deterministic mouse dexterity trainer -- precision, control, and flow.

Built on .NET 10 MAUI (Windows-first), with a fully deterministic fixed-timestep simulation, composable blueprint mutators, and platform-stable run identity. Same seed produces the same level, the same score, and the same replay -- everywhere, every time.

Why MouseTrainer?

• Deterministic to the bit. xorshift32 RNG, FNV-1a 64-bit hashing, fixed 60 Hz timestep with accumulator catch-up. No DateTime.Now, no System.Random, no platform-dependent floats in the hot path.
• Composable difficulty. Six blueprint mutators reshape generated levels before play. Stack them, tune their parameters, and the combination is frozen into the RunId hash for perfect reproducibility.
• Replay verification. Every session records quantized input samples, serialized in a compact binary format (.mtr). Replay verification re-simulates tick-by-tick and checks the event stream hash, score, and combo against the original.
• Modular monolith. Four assemblies with enforced one-way dependencies. Domain is the leaf with zero dependencies; MauiHost is the only composition root. No cycles, no platform leakage into libraries.
• Protocol-grade identity. RunId is an FNV-1a 64-bit hash over mode + seed + difficulty + mutator specs with canonically sorted parameters. Once created, frozen forever.

NuGet Packages

Installation

From NuGet

# Core primitives (RNG, hashing, run identity)
dotnet add package MouseTrainer.Domain

# Simulation engine (game loop, modes, mutators, replay)
dotnet add package MouseTrainer.Simulation

# Audio cue system (event-driven sound)
dotnet add package MouseTrainer.Audio

From Source

git clone https://github.com/mcp-tool-shop-org/MouseTrainer.git
cd MouseTrainer

# Build all library projects
dotnet build src/MouseTrainer.Domain/
dotnet build src/MouseTrainer.Simulation/
dotnet build src/MouseTrainer.Audio/

# Run all tests (214 tests across 10 categories)
dotnet test tests/MouseTrainer.Tests/

# Run MAUI host (Windows -- use Visual Studio, set startup to MauiHost)

Note: The MAUI host project requires Visual Studio with the .NET MAUI workload installed. CLI dotnet build may fail on MauiHost due to MrtCore PRI generation targets -- use Visual Studio for full builds.

Quick Start

using MouseTrainer.Domain.Runs;
using MouseTrainer.Domain.Utility;
using MouseTrainer.Simulation.Core;
using MouseTrainer.Simulation.Modes.ReflexGates;
using MouseTrainer.Simulation.Mutators;
using MouseTrainer.Simulation.Levels;

// 1. Create a run descriptor (deterministic identity)
var run = RunDescriptor.Create(
    mode: new ModeId("ReflexGates"),
    seed: 42,
    difficulty: DifficultyTier.Standard);

// 2. Generate a level blueprint from the seed
var config = new ReflexGateConfig();
var generator = new ReflexGateGenerator(config);
var blueprint = generator.Generate(run.Seed);

// 3. Optionally reshape the level with composable mutators
var registry = new MutatorRegistry();
registry.Register(new MutatorId("NarrowMargin"), 1, spec => new NarrowMarginMutator(spec));
registry.Register(new MutatorId("RhythmLock"), 1, spec => new RhythmLockMutator(spec));

var pipeline = new MutatorPipeline(registry);
var specs = new[]
{
    MutatorSpec.Create(new MutatorId("NarrowMargin"), 1,
        new[] { new MutatorParam("factor", 0.7f) }),
    MutatorSpec.Create(new MutatorId("RhythmLock"), 1,
        new[] { new MutatorParam("div", 4f) }),
};
blueprint = pipeline.Apply(blueprint, specs);

// 4. Wire up the simulation and deterministic loop
var sim = new ReflexGateSimulation(config);
sim.Reset(blueprint);

var loop = new DeterministicLoop(sim, new DeterministicConfig
{
    FixedHz = 60,
    SessionSeed = run.Seed,
});

// 5. Each frame: step the loop with host time and pointer input
// var result = loop.Step(pointerInput, hostNowTicks, ticksPerSecond);
// result.Events contains GameEvent[] for audio, scoring, and UI

Game Modes

ReflexGates

Side-scrolling gate challenge. Oscillating apertures on vertical walls -- navigate the cursor through each gate before the scroll catches you. Deterministic seed produces an identical level every time.

Blueprint Mutators

Six composable transforms that reshape generated levels before play. Applied as an ordered fold over LevelBlueprint:

Mutators are pure functions: LevelBlueprint -> LevelBlueprint. They compose via pipeline (specs.Aggregate), are factory-resolved from MutatorRegistry, and their parameters are frozen into the RunId hash for reproducibility.

SegmentBias Shapes

• Crescendo (shape=0): Easy start, hard finish. d = 2t - 1
• Valley (shape=1): Hard middle, easy ends. d = 8t(1-t) - 1
• Wave (shape=2): Alternating easy/hard segments. d = (-1)^k

DifficultyCurve Exponent

The curve parameter maps to a power exponent via pow(2, curve). Positive values back-load difficulty (easier early, harder late). Negative values front-load it. Zero is identity (no change).

Replay System

Every session can be recorded and verified for anti-cheat and leaderboard integrity.

Wire format: [MTRP magic][Header][RunDescriptor section][InputTrace section][Verification][Checksum]

Audio System

Event-driven audio with deterministic cue selection. The AudioDirector maps simulation events to sound effects with bounded variation -- all deterministic via DeterministicRng.Mix().

Architecture

Four-module modular monolith. No cycles, no platform leakage into libraries.

MouseTrainer.Domain        --> (nothing)          Shared primitives, RNG, run identity
MouseTrainer.Simulation    --> Domain             Deterministic loop, modes, mutators, levels, replay
MouseTrainer.Audio         --> Domain             Cue system, asset verification
MouseTrainer.MauiHost      --> all three          Composition root, MAUI platform host

Prohibited References (Constitutional)

• Audio must never reference Simulation
• Simulation must never reference Audio
• Domain must never reference any sibling module
• No library module may reference Microsoft.Maui.* or any platform SDK
• No mode may cross-reference another mode
• Mutators operate on LevelBlueprint only -- never mode internals

See docs/modular.manifesto.md for the full dependency graph and constitutional rules.

Design Principles

• Determinism is constitutional. Same seed produces the same simulation produces the same score, always. No DateTime.Now, no Random, no platform-dependent floats in the hot path.
• Fixed-timestep simulation. 60 Hz with accumulator-based catch-up. Rendering interpolates between ticks via alpha. Simulation time is derived from tick count (tick * dt), not wall clock, to avoid float drift.
• Protocol-grade identity. MutatorId, ModeId, RunId are permanent -- once created, frozen forever. FNV-1a hashing with canonical parameter serialization ensures the same inputs always produce the same identity.
• Modular monolith, not microservices. Four assemblies with enforced one-way dependencies. Domain is the leaf; MauiHost is the only composition root.
• Warnings are errors. Library projects use <TreatWarningsAsErrors>true</TreatWarningsAsErrors>. MAUI host opts out (SDK-generated warnings). Nullable reference types enabled everywhere.
• Purity in mutators. Blueprint mutators are pure functions with no RNG access, no side effects, and no references to mode-specific types. They read only their parameters and the input blueprint.

Project Structure

src/
  Directory.Build.props         Shared build settings (nullable, warnings-as-errors, analysis level)

  MouseTrainer.Domain/          Leaf module -- events, input, runs, RNG
    Events/                     GameEvent, GameEventType
    Input/                      PointerInput
    Runs/                       RunDescriptor, RunId, MutatorId/Spec/Param, ModeId, DifficultyTier
    Utility/                    DeterministicRng (xorshift32), Fnv1a (64-bit), Leb128 (varint)

  MouseTrainer.Simulation/      Deterministic simulation engine
    Core/                       DeterministicLoop, DeterministicConfig, FrameResult, IGameSimulation
    Debug/                      ISimDebugOverlay
    Levels/                     LevelBlueprint, ILevelGenerator, LevelGeneratorRegistry
    Modes/ReflexGates/          Gate, ReflexGateSimulation, ReflexGateGenerator, ReflexGateConfig
    Mutators/                   IBlueprintMutator, MutatorPipeline, MutatorRegistry, 6 mutators
    Replay/                     ReplayRecorder, ReplayVerifier, ReplaySerializer, InputTrace, EventStreamHasher
    Session/                    SessionController, SessionModels, ScoreBreakdown

  MouseTrainer.Audio/           Audio cue system
    Assets/                     AssetManifest (13 required files), AssetVerifier, IAssetOpener
    Core/                       AudioDirector, AudioCue, AudioCueMap, IAudioSink

  MouseTrainer.MauiHost/        MAUI composition root (Windows)
    GameRenderer.cs             MAUI canvas rendering with neon palette
    GhostPlayback.cs            Replay ghost overlay
    ParticleSystem.cs           Hit/miss particle effects
    ScreenShake.cs              Camera shake on wall hits
    TrailBuffer.cs              Cursor trail rendering
    SessionStore.cs             Local session persistence
    NeonPalette.cs              Color theme

tests/
  MouseTrainer.Tests/           214 tests across 10 categories
    Architecture/               Dependency boundary enforcement
    Determinism/                Replay regression, RNG, session controller
    Levels/                     Generator extraction
    Mutators/                   Blueprint mutator correctness + composition
    Persistence/                Session store
    Replay/                     Serializer, recorder, verifier, quantization, event stream hasher, input trace
    Runs/                       RunDescriptor golden hashes + identity
    Scoring/                    Score breakdown
    Utility/                    LEB128 encoding

tools/
  MouseTrainer.AudioGen/        Audio asset generation tooling

docs/
  modular.manifesto.md          Dependency graph + constitutional rules
  product-boundary.md           Product scope and boundary definition
  MAUI_AssetOpener_Snippet.md   Platform asset wiring snippet

Security & Data Scope

See SECURITY.md for vulnerability reporting.

Scorecard

Full audit: SHIP_GATE.md · SCORECARD.md

License

MIT