Movie Maker Skill

This repo is docs-first: it ships guidance (rules + schemas), not an executable pipeline. This is a skill created with LongStories.ai Skill SEO Audit.

When to use this skill

Use this skill when a user wants a 10-60s mini-movie with a script, shotlist, and render-ready manifest.

What this skill does NOT do

• Does not include a runnable pipeline
• Does not store or upload media
• Does not include provider keys or secrets

What the agent should produce

1. A numbered script (lines L001…)
2. A shotlist divided into scenes (S01…), referencing script lines
3. A manifest JSON that combines both and is ready to render

Start here (MUST do intake)

Read and follow: rules/start/start-here.md Full runbook: rules/start/end-to-end.md

Intake (copy/paste)

Ask these in one message (use an "ask the user questions" tool if available):

1. Length: 30s / 45s / 60s?
2. Aspect ratio: 16:9 or 9:16?
3. Voice: narrator-only (default) or dialogue?
4. Characters: names + 1-2 sentence vibe each (we will lock a canonical descriptor)
5. Music: none (default) or background music?
6. Style: pick a preset (cinematic / 3D animated / anime / 2D / photoreal)
7. Quality mode: Draft (480p), Standard (720p), HQ (1080p)
8. Model preset: Cheap and Fast (Z-Image Turbo + Seedance v1 fast) or High Quality, Expensive (Seedream 4.5 + Seedance 1.5 pro)
9. Reference images: any character/style refs? (if yes, ask user to paste/attach and say you will save to assets/refs/...)

If the user does not answer, propose defaults and ask:

"If I don't hear back, I'll do 30s, 16:9, narrator-only, 3D animated style, Draft (480p), High Quality, Expensive preset. Confirm?"

Core pipeline (voiceover-first)

Script -> Voiceover timing -> Finalize shot durations -> Generate images -> Generate videos -> Render locally -> Share

Model presets (image + video combo)

See: rules/providers/model-presets.md

Quick summary:

• Cheap and Fast: Z-Image Turbo + Seedance v1 fast
• High Quality, Expensive: Seedream 4.5 + Seedance 1.5 pro

Pricing verification

Prices change. Use the fal pricing endpoint in:

• rules/providers/pricing.md

Rules index

Ordered discovery path: rules/index.md

Provider-specific guidance lives under rules/providers/. All other rules should remain provider-agnostic, with defaults listed in rules/providers/defaults.md.

Start

• rules/start/start-here.md
• rules/start/requirements.md
• rules/start/pipeline-voiceover-first.md
• rules/start/end-to-end.md

Planning (script to shotlist)

• rules/planning/script-writing.md
• rules/planning/word-count.md
• rules/planning/scenes.md
• rules/planning/shots.md
• rules/planning/quality-and-cost.md

Manifest

• rules/manifest/schema.md
• rules/manifest/script-and-shotlist.md

Audio (provider-agnostic)

• rules/audio/pauses-and-roomtone.md
• rules/audio/captions-srt.md

Visuals (provider-agnostic)

• rules/visuals/image-prompting.md
• rules/visuals/video-prompting.md
• rules/visuals/style-presets.md
• rules/visuals/characters/character-descriptors.md
• rules/visuals/characters/consistency.md

Timing

• rules/timing/from-timestamps.md
• rules/timing/map-lines-to-timestamps.md
• rules/timing/clamp-and-trim.md

Providers

• rules/providers/defaults.md
• rules/providers/model-presets.md
• rules/providers/fal/overview.md
• rules/providers/fal/api.md
• rules/providers/elevenlabs/voiceover.md
• rules/providers/fal/voiceover-elevenlabs.md (makes ElevenLabs optional)
• rules/providers/elevenlabs/transcription.md
• rules/providers/elevenlabs/music.md

Rendering

• rules/render/rendering.md
• rules/render/assets.md

QA + Troubleshooting

• rules/qa/image-review.md
• rules/qa/video-spotcheck.md
• rules/troubleshooting/common-errors.md

Sharing

• rules/sharing/clawtube.md

Examples

• examples/shipped/ (script + shotlist + manifest)

Advanced

• rules/advanced/overrides.md (per-scene/per-shot model overrides)

Next rules to add

• Provider/model matrix docs
• Safety: secrets, prompt injection, and content policies