claude

Game Design and Build

Make a game, from idea to release. A design studio: ideation, fun targeting, core loop, systems and economy, progression, levels and UX, art direction, a game brief, a prototype plan, design diagrams and a one-page design as HTML, SVG and PDF. A review board backed by cited rules from 38 game-development books: game-studio-cpo and auditors for systems, balance, code, performance and QA. A Godot build kit: scripted bots and balance models, upgrade trees, feature seams and save migrations, headless runs and tests, frame-cost measurement, rendered-frame checks, pixel art, audio, release builds and store pages, and habits for a worker fleet.

1 on GitHubby Matt Scottv2.0.0MITupdated 7 days agogame-dev
/

Plugin game from marketplace skill-marketplace in MattMakes/skill-marketplace. Commands read from the repo's manifest — 6 days ago.

What it installs

Skills (30)

  • art-direction

    Use when a game concept needs its visual direction set: a style sheet (palette, shape language, lighting, texture), a shot list tied to the target fun, and image-generation prompts for a key art, hero and gameplay-mock placeholder. Trigger on: "art direction", "style sheet", "what should this look l

  • balance-model-data

    Keep a game's numbers in a scripted balance model (for example Python modules) that checks design targets and exports the JSON the Godot game loads, and change that data without drift. It covers the edit-export-diff loop, FAIL-key baselines, deterministic exports, parity fixtures between the model a

  • bot-balance-loop

    Tune bosses, encounters and game rules with scripted bots that play the real Godot game headless, then compare the measured tables against written target bands. The loop writes the bands first, builds a fight tool with several bot steerings (coached, tight, wide, wait, ram), runs fixed seeds with re

  • core-loop

    Use when designing or diagramming a game's core loop — applies Stone Librande's "flowchart first" to draw the moment-to-moment, session and long-term loops, each as player action → system response → feedback → next decision, as Mermaid flowcharts, with every feedback edge labeled by the kind of fun

  • design-diagrams

    Use when a game design needs the right diagram, not more prose: an engine loop, a whole-game storyboard, a time-and-space map, a faction relationship chart, an axes-first matrix, or a module map. Trigger on: "what diagram should I use", "map out the loop", "chart the factions", "storyboard the game"

  • feature-seams-and-saves

    Add new content, variants and whole modes to a Godot game through data seams and off-by-default flags, and keep player saves and settings safe across versions. It covers a new playable variant (vehicle, class, character) or encounter batch on a shared base class, an optional row key whose default ke

  • fleet-game-dev

    Habits that let many agent sessions build one Godot game at the same time without breaking each other: one git worktree per plan unit, a fresh import, a baseline, small committed steps, merges of a moving base, a finish gate, and a clean land. It also covers the shared-machine rules (a slot runner f

  • fun-targeting

    Use when choosing what kind of fun a game concept should deliver — picks a proven blend of 2–3 of Marc LeBlanc's eight kinds of fun (Sensation, Fantasy, Narrative, Challenge, Fellowship, Discovery, Expression, Submission) from a curated blend deck, writes a one-sentence experience goal for each, nam

  • game-audio

    Make, wire and measure a Godot game's audio by numbers: seeded procedural sound effects and ambient loops that render the same bytes each time, a file report (length, LUFS, peak, onset, leading silence, loop seam), a bounded audio autoload with tests, a played-mix loudness measure from Movie Maker r

  • game-balance-auditor

    Use when reviewing or tuning GAME BALANCE and ECONOMY — internal economies (resources, currencies, sources/drains), feedback loops, difficulty curves, reward schedules, dominant/degenerate strategies, randomness vs skill, symmetry, or "is this balanced / fair / fun to optimize". Audits the economy a

  • game-brief

    Use when writing up a game concept for a reader — produces the reader-facing brief in a fixed section order: pitch, player fantasy, a 30-second play-by-play, why it's fun (one subsection per target kind of fun), comparable games, scope for a solo developer, risks, the smallest prototype that tests t

  • game-code-reviewer

    Use when reviewing, refactoring, or critiquing GAME CODE / architecture — C#/Unity scripts, systems wiring, class design, patterns, decoupling, state machines, event systems, ECS/data-oriented structure, or "is this codebase well-architected / maintainable". Scores code against a 9-quality rubric an

  • game-ideation

    Use when generating game ideas or concepts — diverges into at least 6 concepts under a constraint card (target fun, core verb, setting, twist, session length, scope), each written through a different ideation frame, rejects any concept that crosses the caller's content boundaries before scoring, sco

  • game-performance-auditor

    Use when reviewing GAME PERFORMANCE / frame budget / optimization — profiling, CPU vs GPU bound, GC and allocations, object pooling, draw calls and batching, culling and LOD, memory and data layout, jobs/ECS, or "why is this slow / will this hit 60fps". Diagnoses the likely bottleneck and proposes r

  • game-qa-reviewer

    Use when reviewing GAME QA, TESTING, or RELEASE-READINESS — test plans and coverage, bug reporting and triage, severity/priority, regression, soak/automated testing, reproducibility, milestone gates and definition-of-done, build/release/certification, version control for releases, or "are we ready t

  • game-studio-cpo

    Use when designing, building, reviewing, critiquing, or making decisions about a game or game system — acts as a veteran studio Chief Product Officer who checks the work against a 2,970-rule studio rulebook distilled and deduped from 38 leading game-dev books. Covers player experience, mechanics & s

  • game-systems-reviewer

    Use when reviewing, designing, critiquing, or upgrading a GAME SYSTEM — an internal economy, progression, combat/damage, loot, crafting, inventory, quest, AI/behavior, save/load, spawning, feedback loop, difficulty/balance system, or the architecture, data, and tooling behind them. Scores a system a

  • godot-perf

    Measure and fix frame cost in a Godot game with numbers you can trust on a shared, noisy dev machine. Covers interleaved A/B pairs against a base build, per-part milliseconds instead of FPS, hitch maps and cost maps, stress scenarios and leak hunts, per-feature cost budget tests, shader warm-up list

  • godot-testing

    Prove Godot changes with a headless GDScript test loop. Record a baseline before you edit, write a test that fails first, run focused filters, then the full suite, a parse check and a smoke scene. Separate new failures from old ones by name. Covers red-first and mutation proofs, flaky tests and shar

  • headless-godot-runs

    Run Godot 4 tools, probes, bots and checks headless so that they compile, end by themselves, report errors honestly and never touch the player's save. It covers `godot -s` SceneTree tool scripts (no class names, autoloads by node path, a quit watchdog), the import and parse check after a merge or a

  • level-ux

    Use when designing a game's levels and player-facing UX — scripts the first 10 minutes beat by beat (what the player does, learns, and which kind of fun it serves), designs exactly 3 sample levels with a layout sketch and difficulty, maps controls for keyboard/mouse, controller and touch, lists the

  • one-page-game-design

    Collaboratively develop or walk through a game or game system, then create a one-page visual design as interactive HTML or detailed standalone SVG with a printable PDF. Headless pipeline mode instead writes a Librande-style one-page JSON spec (title, central illustration, callouts, sidebar, descript

  • pixel-art-pipeline

    Make palette-locked pixel art for a Godot game, from image-model source art or from a drawing script. Each asset gets an exact cell size and a fixed palette, and a check script proves size, palette, frame grid, outline and unchanged old atlas cells. The skill also covers the contrast gate against ev

  • progression-content

    Use when pacing a game from first minute to last — decides the game length up front (Librande), plans at most 10 sessions each with a goal, a new element and a target kind of fun, sets an unlock curve, draws a timed whole-game storyboard of 6–10 panels, and counts the content a solo developer must b

  • prototype-plan

    Use when planning how to build and test a game design — sets exactly 3 milestones (paper prototype, grey-box, vertical slice), each with what it proves, scope, duration and exit criteria; writes playtest questions tied to every target kind of fun; sets measurable success metrics; picks an engine or

  • release-and-store

    Ship a Godot game: export release builds for desktop and Web from the command line, cut a demo or playtest edition from the same project and prove no full-game or dev file leaks into it, make the game fully playable on a gamepad or handheld PC with the right button glyphs, record repeatable gameplay

  • safe-refactor-and-review

    Prove that a code change in a Godot project keeps behaviour unchanged, find every site that enforces a rule before you change it, scan the code base for measured fix candidates, run read-only reviews whose numbered findings small fix units close, and isolate render or feel bugs whose cause is unknow

  • systems-economy

    Use when designing a game's systems and internal economy — lists at most 8 systems and the resources that flow between them as sources, sinks and converters (Adams and Dormans), ties every system and resource to the target kind of fun it serves, sets balance levers with starting values, walks the fi

  • upgrade-tree-design

    Design, price and judge an incremental game's upgrade tree, skill tree or meta-progression from bot and human playtests, not from guesses. It swaps the shipped tree for a dev-only hub-and-spoke diagnostic tree over every real upgradeable stat, lets bots with different buy policies play the real game

  • visual-verification

    Prove that a Godot game looks right with real rendered frames, not headless test logs. The skill stages exact game states in a capture tool, takes before and after images in a real window and at native 1x size, reads every image, measures pixels against a noise floor, runs an automatic windowed visu

View source on GitHub

More from MattMakes/skill-marketplace

All 5 plugins →

Take a change from idea to production, in one plugin. Decide: interview-me, grilling, spec, requirement-convergence. Shape: ddd runs the nine-step DDD Starter Modelling Process with validators, a review page, diagrams and recorded decisions; Blueprint renders architecture, sequence, workflow, data-flow and lifecycle diagrams; ADRs, contracts, module design. Understand: core primes a repo with a knowledge graph and keys keeps one AGENTS.md tree; trace, research-codebase, deepwiki, explain-diff. Build: a planned pipeline from brainstorm to change-description, with tdd, debug, swarm and unlazy. Verify: code-sweep review, complexity, blast radius, verify-feature, harnesses and test suites. Secure: security-sweep. Ship: ship-gate, SLOs, flags, CI/CD, migrations. Principles and stack references are router skills that load one guide at a time. The DDD and Blueprint tools require Node.js 18 or newer. Vendored skills are credited in NOTICE.md.

142 skills · 5 agents

1

Knowledge work, research and decisions, for work that is not code. Decide and research: office-hours, prfaq, plan-ceo-review, roast, deep-research, litreview, deepread, synthesize-research, scientific-problem-selection. Data: profile a dataset, QA an analysis, SQL across warehouse dialects, statistics and A/B tests, Python charts, an offline HTML dashboard. Product operations: metrics reviews, roadmap and stakeholder updates, competitive briefs, market sizing, startup metrics, variance analysis. Support and legal: ticket triage, escalations, contract review, NDA triage. Writing: technical-writing as the entry, with clear-prose and writing-guideline checks, humanize, READMEs and questionnaires. Media and files: HTML slide decks, ffmpeg video editing, file conversion, Mochi flashcard decks and a grounded knowledge vault. Vendored skills are credited in NOTICE.md.

37 skills

1

How should it look and feel? A design-system spine with the design-critic agent: house rules (design-doctrine), DTCG design tokens in three tiers, brand kits, a token build, governance, migration to or from external design systems, Figma sync and DESIGN.md; component, screen and dashboard design with any-framework code generation, image-to-code, redesign and fidelity prototypes; UX writing. Directions against generated-looking UI: frontend-design, impeccable, hallmark, ui-ux-pro-max, Emil Kowalski's design engineering and 138 named design systems. Gates that measure instead of claim: contrast, states, axe, keyboard, target size, overflow, reduced motion and responsive behaviour in light and dark; WCAG 2.2 audits, screen-reader testing and Core Web Vitals. Critique with a Rams lens, design review and plan-design-review. Adapted from plugin87/ux-ui-agent-skills (MIT) and vendored skills; see NOTICE.md.

34 skills · 1 agent

1

Build AI systems. Agents: loop-design-check, agent-architecture-audit, autoresearch-agent, cli-for-agents, writing-for-agents and llm-friendly-context, model-router. Skills and servers: skill-creator, mcp-builder, harvest and book-to-skill. Fleets: herdr-orchestrator, herdr-worker, herdr-atomic and herdr-setup run a multi-agent worker fleet through the horch CLI. Models: decide whether to fine-tune, build the eval harness first, curate datasets, LoRA/QLoRA, preference optimisation, GRPO/RLVR, checkpoint promotion, and RAG and recommender pipeline design. skill-creator and mcp-builder are owned copies of Apache-2.0 skills from anthropics/skills; vendored skills are credited in NOTICE.md.

24 skills

1

Similar plugins

Product discovery skills for PMs: ideation, experiments, assumption testing, feature prioritization, and customer interview synthesis.

13 skills · 5 commands

phuryn/pm-skills 27,939

ideation

claude

Research ideation frameworks including structured brainstorming and creative thinking. Use when exploring new research directions, generating novel ideas, or seeking fresh angles on existing work.

98 skills

Orchestra-Research/AI-Research-SKILLs 14,502

gamedev

claude

Everything: the master router plus all 79 game-dev skills across every engine, discipline, genre, and workflow.

80 skills

gamedev-skills/awesome-gamedev-agent-skills 8,642

adhd

claude

Parallel divergent ideation for coding agents: fans out isolated branches under different cognitive frames, scores, clusters, prunes traps, and deepens the survivors.

1 skill

UditAkhourii/adhd 4,540

Complete YouTube content production workflow: research competitors, generate video ideas, build briefs, craft titles and thumbnails, and create detailed video outlines with demo prep checklists.

5 skills · 7 commands · 3 agents

jeremylongshore/tons-of-skills-marketplace 2,917

Thinking-mode skills that change HOW your AI reasons, not just what it does: escape the generic answer, run cognitive panels, and adversarially stress-test your own decisions. Inspired by parallel-divergent-ideation research.

25 skills

mohitagw15856/pm-claude-skills 1,544

Maintain Game Design and Build?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins5 plugins