一键导入
rpg-illustrator
Generates detailed image prompts and renders images using Nano Banana Pro. Supports Session Recaps, Single Images, and Prompt Generation modes.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Generates detailed image prompts and renders images using Nano Banana Pro. Supports Session Recaps, Single Images, and Prompt Generation modes.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Generates a structured, multi-scene video script optimized for Google Flow (Gemini Omni) from a session recap, maintaining character and location visual continuity via rpg-illustrator prompts.
Generates a narrative session recap from a transcript, using campaign context. Processes the transcript in sequential subagent chunks to preserve chronological order.
Appends session events to the campaign timeline.
Extracts entities (NPCs, Locations, Items, Lore, Handouts) from RPG session recaps and input files, updating the wiki structure.
| name | rpg-illustrator |
| description | Generates detailed image prompts and renders images using Nano Banana Pro. Supports Session Recaps, Single Images, and Prompt Generation modes. |
Use this skill as the central visual engine for the campaign. It handles prompt engineering, style consistency, and image generation, prioritizing defined image_prompt frontmatter from entity files.
This skill accepts the following parameters:
["Session Recap - Prompts Only", "Session Recap - Generate Images", "Single Image", "Prompt Generation"].Session Recap - Prompts Only: Path to the Edited markdown file to process.Session Recap - Generate Images: Path to the session assets folder (content/assets/sessions/{000}/).Single Image: A description string of what to draw.Prompt Generation: A description string of the scene.image_prompt field in its file, COPY IT VERBATIM.
image_prompt, read the entity's file (bio/description) and synthesize a detailed physical description including: race, build, facial features, hair color/style, eye color, clothing, armor, weapons, accessories, and any distinguishing marks.The prompt must be long and detailed.
Follow this structure for complex scenes with multiple characters:
[General Scene Description: Action, Atmosphere, Lighting]
[Character 1 Description (No Name)]
[Character 2 Description (No Name)]
...
[Background/Environment/Secondary Elements]
[Style/Quality Tags]
A chaotic battle scene. In the center, a massive ancient copper dragon with metallic scales and bat-like wings roars, ridden by a spectral figure in obsidian armor conducting with a baton. Facing them are five heroes:
1. young adult male elf with a slender build. He has an angular face with high cheekbones, a sharp jawline, intense dark eyes, and long pointed ears. His shoulder-length, wavy dark brown hair frames his face. He wears layered brown leather armor with intricate silver filigree details over a dark, high-collared tunic. He is equipped with black fingerless leather gloves, an ornate silver pendant with a large cracked dark gemstone, and a single silver diamond-shaped earring.
2. young adult male human wizard with a lean, rugged build. He has intense blue-grey eyes, a strong jawline, a weathered complexion, and a prominent scar across his right cheek. His short, messy dark brown hair is complemented by a full dark brown beard. He wears layered medieval clothing, a rust-brown hooded tunic over a coarse off-white linen shirt, with a thick brown leather shoulder strap and buckle. Two bronze circular geometric star emblems are visible on his chest.
3. powerful male minotaur with a muscular build, broad shoulders, and brown fur. His bovine head features large, curved dark grey horns, piercing glowing blue eyes, and a bull snout. He has stylish, thick blonde hair swept back and a well-groomed blonde fur beard. He is dressed in a draped white toga with an ornate red sash featuring carved patterns and a textured grey shoulder strap.
4. male fantasy warrior with an athletic build and elven features. He wears highly ornate golden plate armor with intricate filigree in a Greco-Roman spartan style. His head is covered by a full golden Corinthian helmet with a large yellow and black plume, from which his glowing orange eyes peer out. He wears a dark blue cloth hood and cape over layered golden pauldrons. His attire is completed by a leather belt with multiple pouches, a gold ornamental bird-head buckle, a blue loincloth with white embroidered patterns, a quiver of arrows, and a round wooden shield.
5. young adult male with a dark fantasy aesthetic. He has a slender build, pale alabaster skin, and short blonde hair styled in a messy quiff. His face is defined by a strong jawline, high cheekbones, and piercing cyan eyes, often held in a confident smirk. large, jagged scar runs along the side of his neck and jaw. He wears a dark blue or black high-collared jacket and a single black leather glove.
In the background, a powerful lizardfolk with a lightning staff and a medusa archer can be seen. The background is filled with cheering spectral spectators. Dramatic lighting, magical effects, intense action. High quality digital fantasy art.
1. hero (Versir): [description]1. hero (Versir): kneeling on the floor, screaming in anger. [description]2. hero (Felicjan): standing in the background, observing silently with crossed arms. [description]Input: Input Data = Absolute path to the session recap markdown file.
Use this mode to generate and save all image prompts without rendering any images yet. Intended as the first phase before user review.
[[wikilinks]] in the text to find relevant entity files for descriptions.### Header in the body, create 2 distinct prompts representing that specific section.content/assets/sessions/{000}/{filename}.txt.
[Prompt](../assets/sessions/{000}/{filename}.txt)
../assets/sessions/{000}/ relative path.generate_image. Return a summary of all saved .txt file paths so the user knows what to review.Input: Input Data = Path to the session assets folder (content/assets/sessions/{000}/).
Use this mode after the user has reviewed and approved the prompts. It reads every .txt file in the assets folder and renders the corresponding image.
.txt files in the given assets folder..txt file, read its content and call generate_image..txt file but with .png extension.Input: Input Data = Description string.
image_prompt or bio, added Style, detailed setting, SOTA structure).generate_image.Input: Input Data = Description description.
image_prompt or bio, detailed setting, SOTA structure).