🕹️ PS42020Sucker Punch ProductionsVERY HARDGhost of Tsushima rebuild deck
VERY HARD
Frame data gives you a duel that tunes cleanly, but a parry only reads as a parry if a hand-authored recovery animation sells it inside four frames, and that is an animator, not a prompt.
You will get a good prototype. The product is what comes after.
A NO here would be dishonest. One duel, four stances and a parry window is a fighting game with a single matchup, and that is a weekend. Fighting-game frame data is the correct model and almost nobody reaches for it in a third-person game, which is exactly why third-person parries usually feel like guesswork. The agent will happily build collider-based hit detection instead; insist on the tables and the debug strip, because every hour of tuning after that is cheap. The thing no prompt reaches is the animation — a parry that does not read inside four frames is not a parry, it is a coin flip with swords — and the island, which is the actual product.
- Difficulty
- VERY HARD · 4 of 5
- First playable
- A project
- The original
- 400 team-years
- Released
- 2020 · PS4
- Genre
- Samurai Action
Build a one-on-one sword duel with a stance system. three.js, or a 2.5D Canvas view. One opponent, one arena, no world. Combat only. Requirements: - Attacks are frame data, not colliders. Every attack is a table of startup, active and recovery frames on a fixed 60Hz tick, and damage lands on a named active frame against a range check. A duel you cannot read frame by frame is a duel you cannot tune. Ship a debug strip along the bottom showing both fighters' current frame and phase. - Parry window of eight frames opening on the defender's input, with a six-frame input buffer in front of it so an early press still counts. The first three frames stagger the attacker; a late parry chips your posture instead. That buffer is the difference between a duel that feels crisp and one that feels broken, and everyone forgets it. - Posture bar separate from health: blocked and parried hits drain it, a break opens a one-hit finisher, and it regenerates only while neither fighter is attacking. - Four stances, each swapping in a different attack table and strong against a different opponent guard pattern. Switching costs twenty frames, so switching mid-combo is a real risk. - Unblockable attacks flagged by a colour flash on the first startup frame and a longer wind-up. Dodge only. - A standoff opener: the opponent approaches, you hold, and releasing on their commit frame is an instant kill. Early loses health. Definition of done: I lose three duels to timing, then win one cleanly on parries alone.
- The art direction, which is the product — wind, pampas grass, colour grading, and a photo mode that flatters every frame
- Hundreds of hand-authored clips; a duel reads because every parry has a bespoke recovery animation
- An island of authored landmarks legible enough to make a marker-free map workable
- Japanese voice performance and a score doing most of the tonal work
Guiding wind — a diegetic direction cue drawn as particles blowing toward your objective, replacing the HUD marker entirely
It moves navigation out of the UI layer and into the world, so the player looks at the scenery instead of a compass. Any game with a destination can use it: dust, leaves, birds, a draught under a door, a dog that trots ahead and looks back.
ScopeAn evening— build just this, not the game
Single self-contained HTML file, Canvas 2D, no build step. Build a top-down walk-around with no HUD and a wind that points the way. Requirements: - A field of scattered trees and rocks, a player dot, three objectives placed off screen. No compass, no minimap, no marker — nothing in the UI layer at all. - A persistent field of 400 grass-blade or leaf particles drifting across the whole screen. Their base direction comes from slow ambient noise; blend that with the unit vector toward the active objective, weighted about 70% toward the objective. - Press W twice to summon a gust: for two seconds the weighting goes to 100%, particle count doubles and speed triples, then decays back. This is the only navigation input in the game. - Grass tufts and tree branches bend to the same vector, so the world agrees with the particles. - Tab switches objective; the field re-converges over about a second rather than snapping. Definition of done: I find all three objectives without ever seeing a UI element.
Not a clone — Ghost of Tsushima re-themed and cut down to something you can finish. The angle is the important reel: it is what turns a project into a weekend. 24,360 ways to land.
Obsidian From Above
Ghost of Tsushima, but a management sim of the same world, set in a Mesoamerican empire. Rendered in VHS camcorder night.
Build a browser game called "Obsidian From Above". The pitch in one line: a samurai action in the spirit of Ghost of Tsushima, but a management sim of the same world, set in a Mesoamerican empire. WHAT THIS IS Take the SHAPE of Ghost of Tsushima — its core loop, and whatever makes that loop good — and rebuild it as your own game in a different world at a smaller scope. Use your own names, art and audio, which you want anyway: a reskin of somebody else's game is less interesting than a new one that works the same way. THE ANGLE — this is the important constraint, honour it above everything else A management sim of the same world. The player never controls a character — they set policy and watch agents with visible needs succeed or fail at it. This is what makes the project finishable. If a decision would push the scope back towards the original's, take the smaller option every time. THE THEME A Mesoamerican empire. Stepped stone architecture, an astronomical calendar that gates events, and obsidian-edged weapons that break. The theme is not a coat of paint. Let it change what the mechanics mean — a reload, a health pack and a locked door should all be things that make sense in this world and nowhere else. LOOK VHS camcorder night. Tracking lines, timestamp burn-in, autofocus hunt, and crushed blacks with over-bright headlights. REQUIREMENTS - Runs in a browser with no build step. Canvas 2D unless the angle genuinely requires 3D, in which case three.js. - Fixed-timestep update loop with interpolated rendering, so it behaves the same at 60Hz and 144Hz. - Every tuning constant in one CONFIG object at the top, exposed as live sliders in a debug panel. You will find the feel by dragging those, not by prompting. - All content — levels, entities, balance numbers — in JSON or plain data files, never inline in the logic. - Audio generated with the Web Audio API rather than asset files. - Respect prefers-reduced-motion: keep the fades, drop the shake and the parallax. BUILD ORDER 1. The core loop with placeholder rectangles, at the reduced scope the angle demands. No theme, no art, no audio. Make it fun as rectangles first. 2. The angle's consequences. Cutting to bots, or to turn-based, or to one level changes the design rather than just shrinking it — find out how and rewrite step 1. 3. The theme, as content and data. 4. The look, applied last as a rendering layer over a game that already works. Start by writing a one-page plan: the core loop, what the angle removes and what that frees you to do properly, and what you are deliberately leaving out. Wait for me to approve it before writing code.
Half the time one of these already does what you wanted, and the other half it saves you a week of solving a problem someone documented in 1997.
One count per game. No account — it is tied to a salted hash of your IP, which is not stored in a form anyone can reverse. Every build adds 400 team-years to the counter on the homepage.
Then try these
Questions
How hard is Ghost of Tsushima to rebuild with an AI agent?
Yes — every game is buildable at some scope. The question is how hard, and Ghost of Tsushima is VERY HARD (4 of 5). You will get a good prototype. The product is what comes after. Frame data gives you a duel that tunes cleanly, but a parry only reads as a parry if a hand-authored recovery animation sells it inside four frames, and that is an animator, not a prompt. Reckon on a project to something you can actually play — not to something that matches the 400 team-years Sucker Punch Productions spent.
What makes Ghost of Tsushima hard to rebuild?
Frame data gives you a duel that tunes cleanly, but a parry only reads as a parry if a hand-authored recovery animation sells it inside four frames, and that is an animator, not a prompt. That is why it sits at VERY HARD rather than a tier either side of it. If you want the idea without the project, the transferable part is guiding wind — a diegetic direction cue drawn as particles blowing toward your objective, replacing the hud marker entirely — an evening of work.
How long does it take to build a Ghost of Tsushima clone?
Wrong question, and it is worth saying why. Nobody rebuilds Ghost of Tsushima — the original took Sucker Punch Productions on the order of 400 team-years, released in 2020 on PS4. What you can do is reach something playable and genuinely yours in about a project, and then keep going for as long as it stays interesting. Any figure that claims otherwise is measuring a prototype and calling it a game.
What do I lose by building my own Ghost of Tsushima?
The art direction, which is the product — wind, pampas grass, colour grading, and a photo mode that flatters every frame. Hundreds of hand-authored clips; a duel reads because every parry has a bespoke recovery animation. An island of authored landmarks legible enough to make a marker-free map workable. Those are the parts a prompt will not hand you.
Which AI coding agent should I use to build it?
Any of Claude Code, Codex or Cursor will handle this prompt. Paste it as your first message, let the agent scaffold the project, then iterate on feel — the second, third and tenth prompts are where a game actually gets good. Work in an empty folder or a fresh git branch so you can throw it away cheaply.
One email when new games go on the list. No other reason to email you, ever.