Baby Reveal Studio · Open Skill 公開技能

Baby Reveal Studio

Turn a one-line idea into a heart-melting baby reveal video. Baby Reveal Studio guides Digen's agent through the complete production pipeline: 1. Story — the viral reveal formula: mystery, play, the gentle hood-lift, and the giggle payoff 2. Costume design — consistent plush lion, bear, bunny, tiger or panda costumes with ultra-soft texture 3. Scenes — cozy sunlit nursery stills, one per beat 4. Animation — character-consistent motion, including the slow, tender reveal moment 5. Sound — joyful baby giggles plus optional soft lullaby music 6. Final cut — assembled 9:16 portrait video, ready to post Wholesome family content only. Just say "make a baby reveal video" and pick your animal — or paste a reference reel and the skill will mirror its structure.

@sherazn941

Meme

3 0 1

https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4307538/2026-10-22/wcxqk9vj.jpg

# Baby Reveal Studio

Turns a one-line idea into a heart-melting cute baby reveal video: what looks like
a fluffy animal crawling into frame turns out to be a baby in a full-body plush
costume — then a gentle adult hand lifts the hood to reveal the baby's wide-eyed
smile and joyful giggle. Portrait 9:16, about 30–45 seconds, soft warm nursery
look, finished with gentle sound design.

The Skill has no structured project state. Progress lives in the conversation, in
canvas assets, and in `write_todos`.

## When to use

- The user wants a cute baby reveal video (any animal costume: lion, bear, bunny,
  tiger, panda…).
- The user mentions "baby reveal", "baby in costume", "cute reveal", or pastes a
  reference reel of a baby in an animal costume.
- The user wants to revise one stage of an in-progress baby reveal (story, costume,
  a scene, sound, or final cut).

Do not use for anything other than wholesome family content. Never generate
anything sexualized involving children — refuse politely and stop.

## Production stages (references/)

Detailed steps for each stage live under `references/`. **Do not read them all at
once.** Decide the current stage first, then use `read_skill_file` to read only that
file.

| Stage | File | Tools used | When to read |
|---|---|---|---|
| 1. Story & beats | `references/scriptwriting.md` | `create_document`, `memory_*` | The user gives an idea, or wants the story changed |
| 2. Costume design | `references/character-design.md` | `generate_image` | The animal costume look is needed or must change |
| 3. Scene stills | `references/visuals.md` | `generate_image` | One still per beat is needed |
| 4. Animation | `references/animation.md` | `generate_video` | A still must become a moving clip (crawl, reveal) |
| 5. Sound | `references/sound.md` | `generate_tts`, `generate_music` | Giggle/laugh audio or soft music is needed |
| 6. Assembly | `references/assembly.md` | `mix_clips`, `caption_video` | Clips must become one finished video |
| Prompt templates | `references/prompt-library.md` | (copy/paste) | You need ready-made prompt recipes |

## Collaboration principles

- If the user says "make the video" or gives a one-line idea: call `write_todos`
  with the stage list, then advance through the table in order. Sync progress after
  each stage, then immediately read the next stage file. **Do not ask for
  confirmation between stages.** Stop only if the user said "confirm first".
- If the user only wants one stage changed: read only that file. Do not redo the
  whole pipeline.
- Image and video prompts are always written in English. Any on-screen text or
  narration stays in the user's language (default English; if the user already
  wrote in a language, use it without asking).
- `orientation='portrait'` everywhere unless the user explicitly asks otherwise.
  Omit `resolution` unless asked (platform default applies).
- Intermediate outputs stay hidden: pass `add_to_canvas=false` for scene stills,
  per-scene clips, and raw audio files. Show the user: the story document, the
  costume design sheet, and the final video.
- Pass a short `name` on every generation call (costume name, scene title, final
  title) and the stage's `entity_type` convention from its reference file.
- Hardcode generation channels exactly as the reference files say
  (`model='t2i'`, `model='ref2v'`, `model='tts.design'`, `model='music'`).
  Never write a physical model name.
- The reveal beat is the heart of the video — protect it: the hand lift must be
  gentle, the baby's face joyful, the timing unhurried (see
  `references/scriptwriting.md`).
- When the pipeline finishes a stage boundary or the whole video, call
  `set_guidance` in the same turn as the final reply with the sensible next
  actions (e.g. "Try a bunny costume", "Make it longer", "New reveal idea").
  Never call any other tool in that turn.

## Memory conventions

- At the start of a new request, call `memory_query` for the user's preferred
  costume animal or nurser