Prompt Details
Model
(claude-5-sonnet)
Token size
955
Example input
[ARTWORK]: A painting showing a seated woman in a rich blue robe, holding an infant on her lap. A halo-like golden circle surrounds her head. The infant reaches one hand toward her face. Background shows a simple architectural interior with a small window
[REFERENCE]: (left blank — tests direct-observation-only mode)
Example output
QUICK READ (2-3 sentences, plain language, zero jargon):
A seated woman wearing a blue robe holds an infant on her lap. The infant reaches one hand toward her face. Behind her head is a golden circular shape, and the setting is a plain indoor space with one small window.
**REFERENCE:** none provided — proceeding on direct observation alone, no sourced grounding to cross-check against.
MOTIF TABLE:
| Element | Category | Confidence |
|---|---|---|
| Seated woman | Natural Element | High |
| Blue robe | Natural Element | High |
| Infant | Natural Element | High |
| Golden circular form behind head | Natural Element | High |
| Architectural interior | Natural Element | High |
| Window | Natural Element | High |
| Woman holding infant on lap | Action | High |
| Infant reaching hand toward woman's face | Action | High |
| Woman's facial expression | Expression | Ambiguous — competes with Natural Element (a facial expression is only "Expression" if a specific visible cue like a downturned mouth or raised brow can be pointed to; without that detail being specified, describing "a face" risks defaulting to a static feature rather than a mood cue) |
AMBIGUITY NOTES:
- Woman's facial expression: the description doesn't specify a visible expressive cue (e.g., eyes, mouth, brow position), so no mood-coloring classification can be supported. Marking this INSUFFICIENT rather than guessing at an emotional tone.
REFERENCE CHECK: not applicable — no reference was supplied for this run.
By purchasing this prompt, you agree to our terms of service
CLAUDE-5-SONNET
Most "describe this painting" prompts jump straight to meaning before
they've actually looked. This one refuses to — and it can ground itself
in real reference material when you have any.
Built on Erwin Panofsky's pre-iconographical level, formalized in the
ICON ontology (Sartini et al., ACM 2023), this prompt enumerates what's
visible BEFORE any interpretation happens. Same retrieval-evaluation
architecture as the rest of this series:
...more
Added 10 hours ago
