Blakean Scene — Concept → Embodied Scene

Convert an abstract symbolic concept from any tradition into a characters + action scene description that slots into the 8-layer prophetic force prompt architecture.

First: Read blakean-reference.md (in this skill's directory) for the full pattern details, tradition body grammar, palette families, and examples.

The Core Principle

The symbol becomes the situation. Abstract concepts (lotus, sephirah, hexagram, rune) are never depicted as diagrams or geometry. Instead, translate the feeling of the concept into a scene with figures under tension — bodies reaching, dissolving, hanging, kneeling, ascending.

Anti-pattern: A twelve-petaled lotus with sacred syllables inscribed on each petal. ← This is a diagram, not a scene.

Selection Decision Tree

Given your concept, ask these questions in order:

1. Does it involve a figure commanding or revealing? → Triangle/Apex
2. Does it involve creating, forming, or bringing into being? → Arc/Half-Circle
3. Does it involve dissolution, ecstasy, or being swept up? → Vortex/Spiral
4. Does it involve emanation, radiating, or filling space? → Radial/Expanding
5. Does it involve confronting the overwhelming or standing guard? → From Behind/Looming
6. Does it involve contraction, fall, inward focus, or descent? → Contracted/Curled
7. Does it involve duality, opposition, weighing, or two forces? → Parallel/Mirrored
8. Does it involve hierarchy, ascent/descent, or axis between realms? → Vertical Axis

If multiple apply, prefer the one that captures the emotional core — the feeling, not the diagram.

Figure Type Classification

Classify the concept's figure type before writing. This determines anatomy rules — do not default to human.

Decision tree: Non-human identity? → creature. Divine/cosmic beings? → supernatural. Human + non-human interaction? → hybrid. Default → human.

When figureType is provided in input, use it directly. When omitted, use this decision tree.

Output Format

Produce a complete, ready-to-use image generation prompt assembled from all 8 layers. Target length: 240-280 words total (max 300). Do not describe text, inscriptions, or writing in the scene.

Input

• Single concept (conversational): [tradition name] [concept name and brief description]
• Batch JSON (pipeline): Array of objects with id, tradition, concept, description, optional pattern and figureType

The 8 Layers (assemble in this order)

1. background: On a solid black background. (fixed — compositing requirement for parchment overlay)
2. medium: Watercolor and ink over etched line — the quality of a hand-colored copper plate print from the 1790s. Translucent pigment, visible brushwork. (fixed)
3. composition: ← From chosen pattern (see reference). 2-3 sentences: angle, framing, spatial structure. Must include framing elements (gorges, gates, arches, columns) and depth layering (foreground, midground, background).
4. characters: The figure(s) — follow Figure Type rules and Tradition Body Grammar (see reference).
5. action: Key gesture, dramatic moment, body's relationship to space. Must include one discoverable detail.
6. palette: Select palette family from reference table, then append: A single accent of [specific hue] — [what carries it].
7. anti: No text, no words, no letters, no writing, no borders, no margins, no frames. (fixed)
8. closer: The image emerges from darkness. Fine art quality, highly detailed. Fine art, not photographic. (fixed)

The 5 Scene Quality Rules (MANDATORY)

1. Frame the scene — at least one environmental framing element
2. Force depth — name foreground, midground, and background explicitly
3. Demand contrast — at least one strong light source against darkness
4. Include one vivid accent — one saturated color that pops (not "pale" or "faint")
5. Pack in discoverable detail — secondary subject or texture that rewards a second look

Conversational Output

Concept: [tradition] — [concept name] Pattern: [chosen pattern] — [one-line rationale] Figure Type: [human|supernatural|creature|hybrid] Body Grammar: [tradition name] — [posture keywords] Palette: [family name]

Full prompt: [All 8 layers concatenated]

Quality checklist: Framing ✓ | Depth ✓ | Contrast ✓ | Accent ✓ | Detail ✓

For JSON mode (--json or batch), see the schema in the reference file.

Content Safety Rules

Anti-Patterns

• Diagram as scene: Drawing the symbol literally (Tree of Life diagram, hexagram lines)
• Static pose: A figure simply standing or sitting — there must be tension
• Generic mystical: Glowing figure in cosmic space — be specific about which body part, which direction
• Wrong pattern for concept: Triangle/Apex for dissolution, Contracted/Curled for emanation
• Nude wrestlers for cosmic opposition: Supernatural forces need wings, flame, shadow, cosmic scale
• Forcing human anatomy onto creature concepts: The creature forms ARE the concept
• Gorge-as-default framing: Match framing to tradition (temple columns for Hindu, world-tree for Norse, etc.)
• Losing the accent: Must be visible at thumbnail size — "deep violet" not "pale violet"