gamestack

Check whether the installed gamestack plugin is behind the marketplace's latest version and update it in place. Use when asked to "upgrade gamestack", "update gamestack", "check for gamestack updates", "is gamestack up to date", "auto update the plugin", "gamestack upgrade", or "what's new in gamestack". Pure plugin-maintenance utility — no design bible, no engine detection, nothing game-specific.

Use when bringing any system — visual (terrain/biome/dungeon/town/castle/character/enemy/NPC/spell-fx), systemic (balance/economy/difficulty), or narrative (lore/quests) — up to fidelity with its intended target through repeated cycles, in any engine. Covers the generic Reference → Diff → Prioritize → Generate/Implement → Verify loop, a preview-harness pattern for screenshotting/inspecting a system in isolation, the animation-can't-be-eyeballed-from-one-frame problem (contact-sheet capture), the external-asset/content-request escalation (when generation is the wrong tool — e.g. animations, trees), the two playtesting channels — human playtesters and an AI playtester that self-verifies via logs/screenshots through a "playtester API" — a human-checkpoint/batching pattern for scaling a generated visual change (including a cheap-mockup gate before heavyweight 3D/tileset generation), and an asset-pack organization + browsable-studio/explorer pattern for what gets approved. Also use to diagnose a project that's gen

The orchestrator for designing a game end to end. Use when starting or steering a game's design — "I'm making a game", "where do I start", "design this game with me", "what's next in the design", or running headless to take a concept to generated, reviewed content. Engine-agnostic; tuned for procedural / AI-authored games (e.g. a procgen open-world RPG in Godot). Sequences the pipeline and names which knowledge/process skill to pull at each phase. Stops at design/spec — does not write gameplay code.

Run this to review a batch of procedurally- or AI-generated game content (dungeons, regions, quests, items, NPCs, lore, history) before it's committed. Stands in for the human designer's eye in a headless generation loop. Runs the oatmeal test (perceptual sameness), the fanfic/retell test (is it worth retelling), a cross-instance sameness scan (structural/thematic duplication), an intentionality gate (who/what/why + completion arc), and an anti-pattern gate. Outputs a structured pass/fail verdict with specific fixes routed back to the generator. Triggers on "review generated content", "check for sameness", "procgen review", "oatmeal test", "is my generated content any good", "gate this content".

Use when designing, choosing, or reviewing the rendering layer of any real-time 3D (or 2D-with-lighting) game — the rasterization pipeline, PBR vs stylized/NPR materials, the forward/deferred/forward+ lighting architecture, global-illumination and shadow choices, culling/LOD/draw-call/overdraw performance budgets, and the post-processing chain (tonemapping, bloom, AO, anti-aliasing). Also use to diagnose a scene that's GPU-bound or CPU-bound, shadow acne or peter-panning, pop-in, overdraw blowout, blown-out or washed-out tonemapping, or content that looks visually incoherent across generated scenes. Triggers on "rendering pipeline", "PBR", "metallic roughness", "deferred vs forward", "forward+", "global illumination", "GI", "lightmaps", "light probes", "SDFGI", "shadow maps", "cascaded shadows", "shadow acne", "peter panning", "frustum culling", "occlusion culling", "LOD", "pop-in", "draw calls", "batching", "instancing", "overdraw", "CPU bound", "GPU bound", "tonemapping", "ACES", "bloom", "SSAO", "ambient o

Use when designing, speccing, or reviewing a character/object motion system in ANY engine — the rig and skeleton, skinning (linear blend vs dual-quaternion, the candy-wrapper artifact), forward vs inverse kinematics, IK solvers (two-bone, FABRIK, CCD, look-at), blend trees and blend spaces (1D/2D), animation state machines, additive layers and upper/lower-body splits, ragdoll and active ragdoll, procedural IK foot placement on uneven terrain, secondary motion (spring bones, cloth, hair), motion matching, animation compression and bone-count budgets, root motion vs in-place, and the anticipation/active/recovery timing decomposition with animation-canceling. Also use to diagnose feet that slide or float, a mesh that pinches when it twists, blends that pop or moonwalk, characters that feel animation-locked, or a motion library that blows the memory budget. Triggers on "animation system", "skeletal animation", "skinning", "candy wrapper", "dual quaternion", "rigging", "inverse kinematics", "IK", "foot placement",

Use when generating or reviewing the GEOMETRY of a game — procedural terrain, meshes, structures, and scatter. Covers noise as the substrate (Perlin/Simplex/value noise, fractal Brownian motion, domain warping), terrain generation (heightfield vs voxel/SDF, marching cubes, dual contouring, hydraulic/thermal erosion, LOD/chunking), structural generation (L-systems, shape grammars, wave function collapse / model synthesis, constraint-based layout), mesh topology correctness (watertight/manifold meshes, inverted normals, winding order, degenerate triangles/UVs, runtime mesh performance and caching), and vegetation/scatter (Poisson-disc/blue-noise placement, slope/altitude/biome masking, GPU instancing). Also use to diagnose terrain that looks like "noise lumps", structures that read as grid-aligned or obviously generated, scatter that clumps or grids, or generated meshes with black/inverted faces and broken collision. Triggers on "procedural terrain", "terrain generation", "Perlin noise", "Simplex noise", "fBm",

Use when authoring shaders/materials or building visual effects in ANY engine — the shader model (vertex/fragment/compute, what each reads and writes), node-graph material thinking, procedural texturing (noise, SDFs, triplanar, vertex-color/splat masks) vs. baked textures, the VFX toolkit (CPU vs GPU particles, trails/ribbons, decals, distortion/refraction, dissolve/hit-flash/rim effects), stylized rendering (toon/cel shading, outline rendering via inverted hull or post-process edge detection), and VFX performance. Also use to diagnose shader compilation stutter, overdraw/fill-rate collapse, an unreadable "particle storm", or generated materials that all look the same. Triggers on "shader", "material", "shader graph", "procedural texture", "noise", "SDF", "triplanar", "particles", "VFX", "visual effects", "trail", "decal", "distortion", "dissolve", "hit flash", "rim light", "fresnel", "toon shader", "cel shading", "outline", "stylized rendering", "shader stutter", "overdraw", "particle storm", "fill rate".

Use when designing, tuning, or reviewing combat and game feel — juice/feedback budgets, hit-stop and impact, telegraphing and danger cues, enemy silhouettes and role-based rosters, multi-enemy readability, encounter pacing, and high-lethality (Souls-like) commitment/stamina/checkpoint loops. Also use to diagnose combat that feels floaty, noisy, unfair, swarmy, or like a memorization "knowledge check". Triggers on "combat design", "game feel", "juice", "hit-stop", "screen shake", "telegraph", "wind-up", "readability", "enemy design", "encounter design", "aggression", "gank", "stamina", "souls-like", "lethality", "this death felt cheap", "combat feels floaty".

Use when designing, tuning, or reviewing the difficulty or balance of any game system — encounter tuning, cost curves, dominant-strategy audits, dynamic difficulty adjustment (DDA), difficulty settings, and accessibility assists. Also use to diagnose "this is too hard/too easy", "everything feels overpowered", a degenerate meta, a missing difficulty curve, or an accessibility gap. Triggers on "balance", "balancing", "difficulty", "difficulty curve", "dynamic difficulty", "DDA", "dominant strategy", "overpowered", "underpowered", "too hard", "too easy", "cost curve", "power budget", "tuning", "accessibility difficulty", "difficulty settings", "assist mode", "rubber band", "flow channel", "tier list", "god tier", "broken build".

Use when designing, tuning, or reviewing the moment-to-moment feel of ANY action in ANY genre — responsiveness, input latency, forgiveness windows (coyote time, input buffering), the juice toolkit (hit-stop, screen shake, squash/stretch, particles, easing, sound layering), the 12 animation principles, camera feel, and UI/menu feel. Also use to diagnose controls that feel floaty, laggy, mushy, weightless, or a screen so noisy you can't read game state. Triggers on "game feel", "juice", "juicy", "feels floaty", "feels unresponsive", "input lag", "coyote time", "input buffering", "hit-stop", "screen shake", "squash and stretch", "easing", "tweening", "feedback", "make it feel good", "game feels mushy".

Use when designing, generating, or reviewing the craft of an individual level, zone, or encounter — layout, wayfinding, structural type, lock-and-key gating, antepieces, arena design, in-level pacing, and the greybox process. Also use to diagnose levels where players get lost, challenge feels unfair, pacing drags, or teaching moments land flat. Triggers on "level design", "level layout", "greybox", "blockout", "wayfinding", "sightlines", "lock and key", "arena design", "encounter design", "the level feels confusing", "players get lost", "chokepoint", "flanking route", "intensity curve", "antepiece", "teach the mechanic", "critical path", "leading lines", "weenie", "breadcrumbs", "level feels boring", "players don't know where to go".

Use when designing, reviewing, or generating the introduction to ANY mechanic in ANY genre — the four-beat teaching pattern (introduce→test→combine→twist), show-don't-tell level design, progressive disclosure and scaffolding, the mechanic dependency graph, FTUE (first-time user experience), early retention funnels, and the skill-atom ramp. Also use to diagnose tutorials that are too long, that players skip, or where players quit before reaching the core loop. Triggers on "onboarding", "tutorial", "tutorialization", "teach the player", "FTUE", "first-time user experience", "new player experience", "players quit early", "players don't understand the mechanic", "tutorial ramp", "skill ramp", "progressive disclosure", "show don't tell", "safe space introduction", "information overload", "tutorial tax".

Use when designing, auditing, or generating the macro arc of an experience — how engagement rises and falls across an encounter, a session, a level, and an entire playthrough. Covers the interest curve and fractal pacing, tension/rest rhythm, novelty cadence and mid-game sag, nested engagement loops and retention ethics, first-hour pacing by genre, and the procedural pacing director algorithm. Triggers on "pacing", "tension curve", "interest curve", "the game drags", "mid-game sag", "player journey", "intensity", "rest beats", "novelty pacing", "retention loop", "the game feels monotonous", "pacing director", "flow state", "pacing director algorithm", "engagement arc", "first hour", "tutorial pacing", "appointment mechanics", "compulsion loop", "safe room", "breathing room", "climax and denouement", "new toy cadence", "encounter pacing".

Use when designing, auditing, or generating a HUD, menu system, or feedback vocabulary for any genre — information hierarchy, diegetic vs. non-diegetic classification, cognitive load budgeting, feedback channel design, menu flow, input modality, and accessibility. Also use to diagnose a UI that is cluttered, that players ignore, that confuses state, or where players miss critical information. Triggers on "UI", "UX", "HUD", "menus", "interface", "diegetic UI", "non-diegetic", "information hierarchy", "cognitive load", "the UI is cluttered", "players miss important info", "feedback vocabulary", "onboarding UI", "accessibility", "colorblind", "button remapping", "contextual HUD", "minimap", "waypoints", "icon soup", "menu hell", "screen shake", "feedback loop", "affordances", "signifiers".

Use when designing high-lethality / permadeath / single-save combat and the systems around it — making death fair and readable, mitigating the cost of failure without removing stakes (short runs, opt-in difficulty, deterministic anti-save-scum), meta-progression and diegetically-integrated death (roguelike vs. roguelite), and the hard open problem of single-save permadeath in a long-form open world. Also use to diagnose deaths that feel unfair, a game churning players at the first big loss, or grind-to-win meta-progression. Triggers on "permadeath", "single-save", "high-lethality", "roguelike", "roguelite", "meta-progression", "save-scumming", "Ironman", "God Mode", "fairness", "telegraphing", "death is progress", "succession", "heir", "Hardcore mode".

Use when building a world bible and deciding how lore reaches the player — the iceberg principle (author the whole world, surface ~10%), environmental & item-description storytelling, an "alien but coherent" identity that survives procedural generation, deep history as faction-proxy pantheons, and ludonarrative harmony (mechanics that are the story). Also use to diagnose lore that info-dumps, feels generic-fantasy, or has gaps that read as contradictions rather than mystery. Triggers on "worldbuilding", "world bible", "lore", "iceberg", "environmental storytelling", "item descriptions", "show don't tell", "alien but coherent", "ludonarrative", "lore dump", "deep history", "mythology".

Use when keeping AI-authored or procedurally-generated content coherent at scale — single voice across a corpus, a machine-readable lore bible with a "never violate" list, generate-then-rationalize causation, recurring thematic domains so generated figures cohere into arcs, the oatmeal/sameness problem, and the self-review pass that flags drift and duplication. The crux skill for headless / AI-authored games. Triggers on "AI-authored content", "keep it coherent", "single voice", "lore bible", "the LLM contradicts itself", "10,000 bowls of oatmeal", "generated content feels samey", "self-review pass", "narrative consistency at scale".

Use when making design-level (not asset-production) decisions about how a game looks and reads — readability vs. fidelity, silhouette design, value/contrast as eye-direction, reserved signal colors, visual-language consistency, and how an AI engine keeps a coherent house style across generated content. Triggers on "art direction", "readability", "silhouette", "signal color", "yellow paint", "visual hierarchy", "style guide", "make it readable", "stylization vs realism", "how should the game look".

The platform router for the gamestack framework. Use when a game's design is ready to implement, when the target engine is chosen or needs choosing, or when handing off a spec to engine code — "build this in Godot/Unity/Unreal/Three.js", "which engine", "now implement it", "wire the design to code". Routes design phases to gamestack's foundation skills and implementation to the matching engine pack (godot, unreal, unity). Engine-agnostic itself; it decides where work goes.

Use when designing quests, reactivity, and factions — quest structure beyond fetch-and-return, branching that honors player choice, reactivity systems on a budget (the "facts database"), procedural/radiant quests as a supplement to hand-authored ones, and faction allegiance dilemmas. Also use to diagnose quests that feel like filler or a world that ignores what the player did. Triggers on "quest design", "fetch quest", "branching", "reactivity", "facts database", "radiant quests", "factions", "choices and consequences", "the world doesn't react to me".

Use when designing, generating, or reviewing the spatial side of an open world — macro-layout and terrain, landmarks and sightlines, biome identity, verticality, navigation/wayfinding, signal color, exploration pull, and spatial pacing (the wonder→fear gradient, tension/release). Also use to diagnose a world that feels empty, samey, or like marker-following "icon vomit". Triggers on "open world", "world map", "world layout", "wayfinding", "triangle rule", "make the world feel alive", "points of interest", "icon vomit", "fast travel".

Use when designing or building a content generator — terrain, dungeons, quests, items, lore, history — or diagnosing generated content that feels samey, generic, or soulless. Covers the "10,000 bowls of oatmeal" problem, perceptual vs. mathematical uniqueness, the handcrafted-anchor + constrained-fill hybrid, voice-consistent corpora, multiplicative systems, and tying generation to meaning. Core skill for headless / AI-authored games. Triggers on "procedural generation", "procgen", "generator", "my generated content is samey", "infinite content", "world generation", "AI-authored content".

The foundational design discipline — the spine the rest of the pack cross-links to. Use when judging whether a mechanic is fun, auditing a system for balance, designing difficulty/flow, setting reward schedules, checking player motivation, or diagnosing "my choices don't matter" / "this is grindy" / "this difficulty spike feels cheap". Covers interesting decisions (Sid Meier), flow & difficulty curves, intrinsic/extrinsic motivation (Self-Determination Theory), feedback & agency, and choice architecture. Triggers on "is this fun", "interesting decision", "dominant strategy", "balance this", "difficulty curve", "flow", "reward schedule", "player motivation", "fake choice", "analysis paralysis", "one more turn".

Use when designing, tuning, or reviewing the core RPG systems triad — character progression (leveling, skills, attributes, classless vs. class), the economy (currencies, faucets/sinks, crafting materials), and loot & itemization (rarity tiers, drop rates, legendary/unique design). Also use to diagnose "leveling feels pointless", "the world scales with me", "players hoard their best items", "every legendary is just +stats", "the economy inflated", or "more loot but it means less". Triggers on "progression", "leveling", "level scaling", "skill system", "attributes", "economy", "currency", "gold sink", "faucet", "inflation", "loot", "itemization", "rarity", "drop rate", "legendary", "build-defining", "auction house", "hoarding".

Use when designing, auditing, or generating the systemic substrate of a game — the small set of consistent, interacting rules (a "chemistry engine") that produce emergent play, rather than scripted set-pieces. Covers immersive-sim principles (intention & perceivable consequence), multiplicative vs. additive design, the "tabletop RPG with a good GM" analogy, and how emergent systems make procedural generation cohere instead of becoming "oatmeal". Also use to diagnose "the world feels inert", "all the variety is scripted", "my generated content is samey", "players can only do what I imagined", or single-solution gates. Triggers on "systemic design", "emergent gameplay", "immersive sim", "multiplicative gameplay", "chemistry engine", "affordances not solutions", "intention and consequence", "interacting systems", "oatmeal", "every quest is the same".