Generates or edits images on the filmmaking canvas via the local generateimage.js CLI, following canvas-aware conventions for subtype stamping, reference wiring, and provenance edges. Use when the user asks to design a character, portrait, location, hero still, variation, edit, scene shot, multi-view reference sheet, or storyboard mosaic. CHARACTER DESIGN ROUTING (read first) — Whenever a character will appear in downstream video work (anything the user calls a video / clip / promo / 宣传片 / 短片 / 连续剧 / film / scene / 拍片 / shot — even if the canvas is empty and the character is being designed from scratch), DEFAULT to Pattern 7 (4-panel character reference sheet — Front/Profile/Back/Closeup), NOT Pattern 1 (single portrait). This holds whether the user has uploaded actor reference photos or not. With ≥3 actor refs, Pattern 7 uses multi-ref triangulation; with 0-2 refs, Pattern 7 uses text-only generation but still emits the 4-panel layout, because the video model needs multi-view anchor data to keep the character recognizable across non-front shots. Pattern 1 (single portrait) is only for one-off static stills that will NOT feed video gen — character posters, print art, standalone illustrations. Before any character generation, briefly announce the choice to the user ("Starting with a 4-panel reference sheet for each character so the identity stays locked across video shots — tell me if you want simple portraits instead"). Never silently default to Pattern 1 when video work is implied. STORYBOARD SHAPE — storyboard / mosaic / 分镜 / shot list / keyframe sheet / coverage / previs = ONE composite image with N×M panels per location, NOT N separate calls. Re-invoke this skill (Pattern 6) before firing the CLI.
Install with
Claude CodeCodexCursorGitHub CopilotGemini CLI
Supported by the Atlas install prompt. Skill-specific requirements vary.
Skill stats
Skills1CategoryGeneralStars0
About this skill
Generates or edits images on the filmmaking canvas via the local generateimage.js CLI, following canvas-aware conventions for subtype stamping, reference wiring, and provenance edges. Use when the user asks to design a character, portrait, location, hero still, variation, edit, scene shot, multi-view reference sheet, or storyboard mosaic. CHARACTER DESIGN ROUTING (read first) — Whenever a character will appear in downstream video work (anything the user calls a video / clip / promo / 宣传片 / 短片 / 连续剧 / film / scene / 拍片 / shot — even if the canvas is empty and the character is being designed from scratch), DEFAULT to Pattern 7 (4-panel character reference sheet — Front/Profile/Back/Closeup), NOT Pattern 1 (single portrait). This holds whether the user has uploaded actor reference photos or not. With ≥3 actor refs, Pattern 7 uses multi-ref triangulation; with 0-2 refs, Pattern 7 uses text-only generation but still emits the 4-panel layout, because the video model needs multi-view anchor data to keep the character recognizable across non-front shots. Pattern 1 (single portrait) is only for one-off static stills that will NOT feed video gen — character posters, print art, standalone illustrations. Before any character generation, briefly announce the choice to the user ("Starting with a 4-panel reference sheet for each character so the identity stays locked across video shots — tell me if you want simple portraits instead"). Never silently default to Pattern 1 when video work is implied. STORYBOARD SHAPE — storyboard / mosaic / 分镜 / shot list / keyframe sheet / coverage / previs = ONE composite image with N×M panels per location, NOT N separate calls. Re-invoke this skill (Pattern 6) before firing the CLI.
Quick start
Copy the install prompt. Use the green button above this README.
Send it to your agent. Paste it into your coding agent’s chat and send.
Review the setup. Your agent checks the skill’s instructions, requirements, and installation scope.
Generates or edits images on the filmmaking canvas via the local generate_image.js CLI, following canvas-aware conventions for subtype stamping, reference wiring, and provenance edges. Use when the user asks to design a character, portrait, location, hero still, variation, edit, scene shot, multi-view reference sheet, or storyboard mosaic. CHARACTER DESIGN ROUTING (read first) — Whenever a character will appear in downstream video work (anything the user calls a video / clip / promo / 宣传片 / 短片 / 连续剧 / film / scene / 拍片 / shot — even if the canvas is empty and the character is being designed from scratch), DEFAULT to Pattern 7 (4-panel character reference sheet — Front/Profile/Back/Closeup), NOT Pattern 1 (single portrait). This holds whether the user has uploaded actor reference photos or not. With ≥3 actor refs, Pattern 7 uses multi-ref triangulation; with 0-2 refs, Pattern 7 uses text-only generation but still emits the 4-panel layout, because the video model needs multi-view anchor data to keep the character recognizable across non-front shots. Pattern 1 (single portrait) is only for one-off static stills that will NOT feed video gen — character posters, print art, standalone illustrations. Before any character generation, briefly announce the choice to the user ("Starting with a 4-panel reference sheet for each character so the identity stays locked across video shots — tell me if you want simple portraits instead"). Never silently default to Pattern 1 when video work is implied. STORYBOARD SHAPE — storyboard / mosaic / 分镜 / shot list / keyframe sheet / coverage / previs = ONE composite image with N×M panels per location, NOT N separate calls. Re-invoke this skill (Pattern 6) before firing the CLI.
About this listing
Category
General
Reported license
Not specified
Listing language (estimated)
English
Review status
Not individually reviewed
Agent compatibility
Atlas installation supports Claude Code, Codex, and Cursor. Skill-specific dependencies and features are checked during setup.
Claude Code
Codex
Cursor
GitHub Copilot
Gemini CLI