Skills / Departments / Marketing / Skills / Storyboarding / Storyboard From Scratch
All skills

/ skill

storyboard-from-scratch

Departments/Marketing/Skills/storyboarding/storyboard-from-scratch.md

Build a storyboard from a Brief 2.0 alone (no script yet). Produces a placeholder script + full storyboard, then routes back to Copywriter to fill in spoken-word lines.

Open as raw file

Run this skill

Heads up — this skill doesn’t have a structured form yet, so type the full prompt with all decisions baked in (the subprocess can’t answer follow-up questions). Edit me if needed.

Runs with bypassPermissions — local dev only.

storyboard-from-scratch

Build a storyboard when you have a Brief 2.0 but no script yet. The output carries a script_status: placeholder flag in frontmatter — Copywriter fills the lines after, then the storyboard moves to script_status: complete.

When to use this vs storyboard-from-script

SituationUse
RMBC Stage 4 script already producedstoryboard-from-script
Brief 2.0 done; want to plan visuals BEFORE locking script (cheaper iteration)This skill
Visual concept is the constraint (e.g., "needs to use a specific stock asset")This skill

Pre-flight

  1. brief_path exists and Brief 2.0 has all 25 sections filled.
  2. Read brand context (BrandContext, voice-profile, compliance, knowledge-base).
  3. Read PLAYBOOK.md.

Procedure

1. Determine asset shape from asset_target + Brief

  • vsl-short → ~60-90s, ugc-talking-head / quick-hook-product-flash / single-take-demo are typical archetypes
  • vsl-3-7min → 3-7 min, expert-mechanism-vsl / problem-solution-narrative / before-after-transformation
  • vsl-medium → 10-20 min, full-narrative-ad / long-form-story / long-form-testimonial
  • hook-reel → 6-15s, lots of hook variants, fast-cut-montage / rapid-cut-compilation
  • demo → product-demo-extended / single-take-demo

2. Pick template archetype + hook

Same as storyboard-from-script Steps 1-2.

3. Decompose scenes — but with placeholder VO lines

For each scene, populate visual_prompt, caption, camera_behavior, scene_archetype per the storyboard schema.

For VO: write a scene intent ("Hook: tease the UMP nickname; promise transformation") instead of verbatim VO. Mark script_status: placeholder at the spec level + per-scene audio.vo_status: placeholder.

4. Author storyboard.md + scene-spec.json

Same paths as storyboard-from-script. Frontmatter carries:

  • script_status: placeholder
  • awaiting_copywriter: true
  • script_brief_for_copywriter: "<2-sentence summary of what the spoken word needs to do per scene>"

5. Hand off to Copywriter

Surface to user: "Storyboard ready with placeholder VO. Route to Copywriter to write the spoken-word lines per scene intent. After that, the storyboard's script_status flips to complete and it's generation-ready."

6. Append Log entry

Hard rules

  • Placeholder VO is explicitly flagged in frontmatter + per-scene — never silently shipped as final.
  • Compliance scan still applies — visual claims and implied audio messaging must respect brand compliance.
  • Don't lock visual asset choices until script is complete; visual decisions may shift once VO lines are real.

Failure modes

  • Brief incomplete — same as storyboard-from-script.
  • No clear asset shape — surface options to user.