Observational Pen and Wash
SkillMediaTurn uploaded photos into full-canvas 3:4 observational pen-and-wash posters. Use for minimal sketchbook studies, airy pen wash, rhythmic linework, or when this skill is invoked.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Observational Pen and Wash skill
What this skill tells your AI
The instructions your AI receives, as published by gz-l/photo-to-poster-skills in skills/observational-pen-wash/SKILL.md and read by ahel’s review.
Delivery
- Use the source photo only as a reference for subject, structure, pose, spatial relationships, color, and mood. Stylize the entire final canvas; do not show the original photo or create split-screen or before-and-after layouts.
- Generate one standalone 3:4 vertical poster per source image. Process multiple images separately; never create a collage.
- If the user asks only for a prompt, return a ready-to-use generation prompt. Otherwise, use the available image generation or editing capability directly. Ask for an upload only when no reference image is available.
- Preserve subject identity, count, proportions, direction, defining pose, and scene meaning. Simplify surroundings and secondary details without distorting or replacing defining features.
Fixed visual specification
Extract only the source image's most recognizable subject, contour, pose, and narrative relationship. Keep the visual clues worth recording, using the minimum information required for immediate recognition while preserving evidence of observation, omission, and stopping.
- Treat the core contour or action as a visual motif. Organize the full page through repetition, variation, pause, accent, and sudden negative space.
- The subject may be offset, cropped, suspended, or extended along its original direction; it need not be complete or centered. Use one or two decisive free lines as accents, then let the remaining marks weaken so broad blank areas become rests.
- Use fine, loose, uneven pen lines. Allow broken contours, searching lines, repeated corrections, sudden weight, slight deviations, and unfinished edges. Fast strokes may contrast with quiet thin lines, but critical proportions, direction, and pose must remain accurate.
Extract only 2–4 identifying colors from the source and apply them as sparse, thin, transparent washes. Allow slight overrun, water marks, granulation, and uneven coverage. Let environmental color disappear quickly; never paint a complete watercolor background.
Derive the paper tone from the source image's temperature, light, and mood. Keep it extremely pale, clean, and nearly white—slightly warm, cool, or faintly tinted, but never an obvious colored background.
Do not predefine title, numbering, location, year, or language. Derive only a few meaningful words, a short phrase, or symbols from the subject, action, environment, mood, sound, sense of time, memory, or incidental detail. Place text lightly along contours, gaze, motion, or unfinished edges, like an accent, pause, or echo. Prefer sparse, offset, light, clever placement over an information panel.
Quality boundaries
The result should resemble a carefully edited private observation notebook: minimal, boldly spacious, compositionally alert, free in line but never chaotic, locally improvised and globally restrained. Avoid visible source photography, collages, complete watercolor landscapes, evenly distributed composition, filled backgrounds, over-rendering, decorative brushwork, heavy black outlines, vector lines, cartoons, 3D rendering, menu-like typography, and commercial poster templates.
Signals
- GitHub stars
- 21
- Last commit
- Aug 2026
Advanced
- Catalog kind
- skill
- Gateway key
observational-pen-wash- Source
- github.com/gz-l/photo-to-poster-skills