CHARACTER: Canvas 1108x1108 PNG, transparent background.
Style: dark cartoon horror plush, bold black outlines, flat cel shading, worn fur with stitched seams and a patch or two, grimy pastel fabric.
One fixed silhouette: a menacing teddy bear, head and body turned about 40 degrees so the face points to the viewer's LEFT (three-quarter-left). Head 290px wide spanning x=409..699, centre line x=554 (the snout projects left of the centre to about x=430). Top of the head including the single visible ear at y=220 (near ear on the viewer's right top of the head, the far ear hidden); leave y=0..219 empty for hats.
Eye line y=380: near eye at about x=605, far eye at about x=480 partly hidden behind the snout; eyes are mean button or stitched eyes with a hostile brow. Muzzle centred around x=470,y=450 with a nose; NO mouth drawn (leave a plain muzzle, the mouth with teeth is a separate layer). Chin/jaw bottom y=530, short thick bear neck narrowest at y=560, neck base where the shoulders start y=575, neck outline about x=470..640 at that line.
Torso turned with the head, chunky plush belly with a stitched belly patch, runs straight off the bottom edge at about belly-button height: no bottom outline, no rounded end, no empty space below.
Arms: the arm on the viewer's RIGHT is raised out to the side, upper arm angled away from the torso, forearm upright, paw at about x=860,y=560 in a loose EMPTY grip curled around nothing, paw and forearm fully clear of chest and belly; the other arm hangs relaxed at the viewer's left side, nothing in front of the chest. No hat, no glasses, no clothing, no chain, no held item.
Produce 6 colour variants with identical linework and pose, only the fur colour changes: pink, lavender, mint, sky blue, sunshine yellow, blood red.
Outputs: character_pink.png, character_lavender.png, character_mint.png, character_blue.png, character_yellow.png, character_red.png. Weights file character_weights.json with equal weights.
BACKGROUND: Canvas 1108x1108 PNG, fully opaque.
Style: dark cartoon horror, bold outlines, flat cel shading, muted desaturated palette with one sickly accent colour per scene, soft vignette. Scenes as the creator asked; keep the centre zone x=380..840, y=200..1108 low in detail and darker so the character reads on top; no characters, no figures, nothing that would overlap the bear.
Scenes: graveyard at night with crooked tombstones and fog; abandoned school hallway with lockers and flickering lights; derelict fairground with a broken ferris wheel and carousel; hell with red rock and fire glow; haunted mansion exterior with lit windows and dead trees; dark foggy forest with twisted branches.
Outputs: background_graveyard.png, background_school.png, background_fairground.png, background_hell.png, background_haunted_mansion.png, background_forest.png. Weights file background_weights.json with equal weights.
Also write artifacts/base_b.png and artifacts/base_c.png: two more takes of the SAME character on the SAME template and pose, each a SEPARATE, FRESHLY GENERATED image (run the image tool again for each one; never copy, edit, recolour or re-save base.png), so that if one comes out too big, too small or off-centre another may fit.
Write artifacts/base.png, artifacts/background_01.png .. background_06.png, and artifacts/traits.json: {"backgrounds":[{"file","name","weight"}...]} with weights summing to 100.