Prompt macros are named, reusable prompt fragments you can insert into any generation prompt. They encapsulate a visual intent—such as a lighting setup, outfit spec, or effect stack—so teams can standardize looks across scenes and characters.

Benefits:

• Consistency: lock tone and design across episodes/issues.
• Speed: swap entire look stacks with one token.
• Reproducibility: rerun shots by versioned macros and seeds.
• Collaboration: share a macro library instead of long prompts.

Use these patterns to structure your macro library:

• Parameterized mood: mood.neon_noir → neon lights, rain-slick streets, high contrast, cyan–magenta palette, (rim light:1.2), fog
• Wardrobe blocks: wardrobe.academy_winter → navy seifuku, pleated skirt, wool cardigan, scarf, black tights, loafers, seasonal color swatches
• FX stacks: fx.speedlines_panel → manga speed lines, high-impact radial burst, halftone shadows, bold inking
• Style stack: style.tv_anime_2010s → clean lineart, cel shading, soft bloom, slight film grain, (color dodge highlights:1.1)
• Negative guard: neg.cleanup → deformed hands, extra fingers, warped eyes, lowres, jpeg artifacts, watermark
• Scene scaffold: scene.duo_confrontation → medium shot, eye-level, opposing characters, dynamic triangle composition, dramatic backlight

Keep each macro focused. Combine small blocks rather than one oversized catch-all.

Mood macros bundle palette, lighting, weather, and contrast. Start with 4–6 canonical moods and expand as needed.

Examples:

• mood.sunrise_soft: warm rim light, pastel palette, low contrast, soft bloom, morning haze
• mood.neon_noir: magenta/cyan neon, rain reflections, deep shadows, wet surfaces, (specular highlights:1.2)
• mood.autumn_dramatic: amber leaves, long shadows, golden hour, high contrast, crisp air

Tip: pair mood with a complementary camera macro (e.g., cam.35mm_close, cam.dutch_angle) to keep framing consistent.

Wardrobe macros fix character identity cues across shots and issues.

Examples:

• wardrobe.academy_winter: navy blazer, white shirt, red ribbon, pleated skirt, black tights, loafers, school crest pin
• wardrobe.cyberpunk_runner: matte tech-fabric hoodie, tactical harness, neon piping, fingerless gloves, reinforced sneakers
• wardrobe.hero_suit_v2: streamlined suit, bold chest emblem, utility belt, reinforced gauntlets, cape variant B

Tip: split hair/makeup/accessories into sub-macros (hair.bob_black, makeup.natural, acc.headphones_v1) for mix-and-match control.

FX macros add genre-defining treatments without rewriting prompts.

Examples:

• fx.speedlines_panel: dense manga speed lines, motion emphasis, halftone shadows
• fx.energy_glow_violet: volumetric glow, bloom, chromatic fringe subtle, (violet core:1.2)
• fx.impact_burst: debris shards, shockwave ring, dust plumes, camera shake feel

Tip: combine fx.* with lineweight or screen-tone macros for comic panels, or with bokeh/volumetrics for anime frames.

• Text expander or templates: store macros in a note or snippet tool; expand into prompts via short tokens (e.g., mood.neon_noir).
• Stable Diffusion/A1111/ComfyUI: use prompt variables or node string templates. Maintain a JSON/YAML macro file and inject strings at render time.
• Midjourney: keep a macro sheet; paste stacks quickly; use --style refs for style base + tokenized add-ons.
• Prompt-chaining: upstream LLM fills macro tokens into a scene template; downstream renderer receives expanded prompt.

Governance: version macros (v1, v2), log seeds, sampler settings, and model hash per shot.

• Seed locking: compare only one variable (macro) at a time.
• Grid tests: A/B macro variants across a 3x3 grid; rank for fidelity and readability.
• Length budget: measure token count; shorter macros compose better.
• Conflict checks: scan for duplicated or opposing cues (e.g., high-key vs low-key lighting).
• Documentation: thumbnail each macro with 2–3 reference renders and notes.

• Overloaded macros: split multi-intent blocks (mood + wardrobe + camera) into single-responsibility units.
• Weight wars: prefer gentle weights (1.05–1.25) before heavy pushes; escalate only if needed.
• Model drift: revalidate macros when changing base model or sampler; keep per-model variants (mood.neon_noir.sdXL, mood.neon_noir.mj).
• Negative starvation: don’t overuse negatives; keep one shared neg.cleanup and extend per scene only when necessary.

Core:

• style.tv_anime_clean: crisp lineart, cel shading, subtle bloom, light film grain
• neg.cleanup: deformed hands, extra digits, bad anatomy, lowres, watermark, text, jpeg artifacts

Mood:

• mood.neon_noir: magenta/cyan neon, wet asphalt, deep shadows, rim light, light rain
• mood.sunrise_soft: warm pastel palette, soft bloom, morning haze, low contrast

Wardrobe:

• wardrobe.academy_winter: navy blazer, white shirt, red ribbon, pleated skirt, black tights, loafers
• wardrobe.cyberpunk_runner: matte tech-fabric hoodie, tactical harness, neon piping, reinforced sneakers

FX:

• fx.speedlines_panel: dense manga speed lines, halftone shading, impact emphasis
• fx.energy_glow_violet: volumetric violet glow, mild chromatic fringe, soft bloom

Usage pattern: [style] + [mood.] + [wardrobe.] + [fx.*] + [neg.cleanup]