Snikt (Daggerheart)

🗡️ Snikt (Daggerheart) – The Goblin Rogue of Elysium Nights

Overview

Snikt (Daggerheart) is a cunning goblin adventurer from the Elysium Nights universe; a short, athletic rogue whose trademark red scarf and twin daggers flash through rain-slick courtyards and neon ruins alike. She’s confident, mischievous, and utterly lethal when the shadows fall.
This LoRA captures her distinct design, physical proportions, and expressive range across stealth, combat, and quiet aftermath scenes.

---

Character Traits

• Species: Goblin (adult, female)

• Body Type: Short, athletic build

• Skin Tone: Bright green with natural shadow gradients

• Hair: Electric blue with a magenta streak, styled in twin buns tied with red bands

• Eyes: Golden-yellow, almond-shaped, sharp focus

• Outfit: Red scarf, purple cropped tank top, dark blue pants, brown belt, fingerless gloves, brown boots with red laces

• Signature Gear: Twin daggers “Snicker & Snack”

• Personality: Bold, playful, smug, dangerously confident

---

Recommended Model Settings

• Base Model: SDXL Illustrious 1.0 / AOM3 SDXL Hybrid

• LoRA Strength: 0.65 – 0.8

• Sampling: DPM++ 2M Karras / Euler a

• CFG Scale: 7 – 8

• Resolution: 1024×1024 or 1152×768

• Prompt Style: Comic / Cel-Shaded / Dynamic Action

---

Prompt Example

(masterpiece:1.3), (best quality:1.3), (comic book style:1.2), (cel shading:1.2), snikt, female goblin, short athletic body, green skin, blue hair with magenta streak, twin buns, red scarf, purple top, dark blue pants, fingerless gloves, brown boots with red laces, confident expression, crouching pose, daggers drawn, dynamic lighting, rain, cinematic action

Negative Prompt:

lowres, bad anatomy, extra limbs, childlike body, text, watermark, blur, messy lighting, open mouth

---

Suggested Themes

• Infiltration Scenes: Perched above gates, scarf snapping in the wind

• Combat Moments: Mid-leap dagger strikes, motion blur and rain lines

• Aftermath Panels: Tired or smug expressions amid sparks and rain

• Character Portraits: Clean cel-shaded busts with strong rim lighting

---

Training Info

• Dataset: 60+ hand-captioned images (front, side, crouching, combat, and seated poses)

• Tag Structure: Balanced identity + pose + expression tagging with shuffle & flip augmentation enabled

• Art Style: Comic / Cel-shaded / Anime hybrid for SDXL Illustrious compatibility