Cinematic Wilds Deck

Provide topic, audience, workshop purpose, approved messages/story material, language and exact count. Supply evidence for any real-world claim. Agree a missing count from content and duration, counting opening/closing unless specified otherwise.

Plan page jobs from world/place/character to framing, sequence and detail, selecting or extending the twelve roles. Preserve the black/white/cyan/lime system and the three-font hierarchy. The themes reference gives one authored direction, not multiple editor presets.

Create an art ledger per selected slot with subject, aspect, placement, palette, prompt and result. Choose suitable library reuse, reuse with generated gaps or regenerate-all, using approved supplied assets first. Twenty-one local scenes are available, but fictional animal-story assets do not become documentary evidence for another subject.

Match wide landscapes, 4:3/4:5 character fragments, tall 3:4 film strips and square detail/lens scenes to the actual kit slots. Use each artwork once; a crop/recolor/rotation is still the same source. Preserve simplified expressive faces, angular painted planes and matte fur, without copying known movie characters or introducing photoreal animal anatomy as a style substitute.

Embed catalog IDs through the checksum-verifying Node helper. Authorized new artwork may use a built-in tool or optional paid FAL path with FAL_KEY, recording returned task IDs for status/fetch without automatic failure retries. Costs/account limits may apply. When a prompt asks to edit a prior result, use only suitable approved/prior deck art or generate from text; list unresolved frames.

Replace fictional quotes, narrative, scene descriptions, exercise labels and notes with approved material. Never fabricate attribution, film affiliation or wildlife observations. Keep words/prompts/diagram shapes editable outside artwork. Compose/build/check with Python 3 and the agreed count/style kit; Node serves media helpers. Embed assets below 20 MB and clear sample leftovers.

Open over local HTTP and inspect all 1280×720 pages for expressive-title readability, film-strip crops, quote attribution, prompt contrast, unique images and bounds. Confirm counters/thumbnails/notes match the final list and test Edit/E, text/images/order/notes/object changes, undo/redo, Save and reopening.

Export requested PPTX/PDF, run the PPTX checker and inspect text/notes, rendered painted scenes/diagrams and actual fonts in the target application. Report actual visual acceptance separately from source-fact validation; a scene called motion remains a presentation page unless another output is explicitly built.