skills.rest
TeamWiseFlow
TeamWiseFlowOFFICIAL
Explore all skills by this author →
GitHub

collage-broll

Generates editorial halftone paper-collage B-roll videos from short voiceover scripts via staged approval gates.

8.5k|1.4k|Updated Apr 24, 2024
Search Tags:#FFmpeg#video generation#image-to-video#b-roll#short video#Storyboarding#paper collage
One-click install
npx skills add https://github.com/TeamWiseFlow/wiseflow --skill collage-broll
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: collage-broll
Source: https://github.com/TeamWiseFlow/wiseflow/tree/main/crews/content-producer/skills/collage-broll
Command: npx skills add https://github.com/TeamWiseFlow/wiseflow --skill collage-broll

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Turning a 5-second voiceover line or abstract concept into a visually coherent B-roll clip normally requires manual storyboarding, image generation, and video assembly. This Skill automates that pipeline while preventing wasted video-generation spend through mandatory human approval checkpoints.

Core Features & Use Cases

  • Three-Gate Approval Workflow: Proposes visual metaphors first (Gate 1), generates color collage stills via siliconflow-img-gen/Seedream only after confirmation (Gate 2), then produces video via aigc-video-gen i2v first/last-frame interpolation (Gate 3).
  • Consistent Editorial Style: Enforces a unified design language—flat bold color fields, black-and-white halftone cut-outs, colored cardstock accents, cream keylines, and assemble-from-empty motion—across batched clips.
  • Automated QA and Delivery: Builds contact sheets with ffmpeg, compares video end frames against approved stills, runs the video-review technical gate, and delivers silent 9:16 720x1280 5-second MP4s by default.
  • Use Case: A content creator pastes five voiceover lines for a short video; the Skill proposes one visual metaphor per line, generates approved collage stills, and outputs five ready-to-use vertical B-roll clips.

Quick Start

Ask the agent to turn this voiceover script into collage-style B-roll clips using the collage b-roll workflow, then approve the proposed visual metaphors and still frames at each gate.

Frequently Asked Questions about collage-broll

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I turn a voiceover script into B-roll video clips?▼

Provide the script and the Skill proposes one visual metaphor per line for approval, generates collage still frames with siliconflow-img-gen after confirmation, then calls aigc-video-gen in i2v mode to animate each approved still into a 5-second clip.

What video generation models does the collage B-roll workflow use?▼

Video generation goes through the aigc-video-gen wrapper, which defaults to Bailian happyhorse-1.1-i2v with fallback to 1.0 and wan2.7, or Volcano Seedance Fast/Normal/Mini when Bailian keys are not configured. Still images use Seedream via siliconflow-img-gen.

What API keys and tools are required to run this workflow?▼

You need AWK_API_KEY for still image generation, plus MODELSTUDIO_API_KEY/DASHSCOPE_API_KEY for Bailian video or AWK_GEN_KEY for Volcano Seedance. ffmpeg, ffprobe, and Python 3.10+ must be installed; scripts/check_setup.sh verifies all of these.

Why does the i2v video generation step fail with a non-zero exit code?▼

Check that first-frame.png and last-frame.png actually exist and are exactly 720x1280, and that the working directory is the workspace root. The aigc-video-gen wrapper requires relative output paths under output_videos/ or a platform outputs directory.

When should I not use the paper-collage B-roll approach?▼

Avoid it when you need precise layer control, editable timelines, camera moves through scenes, real actors, or product advertising footage. For layer-by-layer editable animation, use a layered animation tool such as Manim instead.

Does the final collage video include audio?▼

By default the Skill delivers a silent version, final-5s-noaudio.mp4, stripped with ffmpeg after generation. If you explicitly request audio, it delivers the original final-5s.mp4 with the sound effects and BGM produced by the video model.

Related Skills

Explore More Content & Communication Skills→
brudexbrudex

auto-video-generation

Automates video briefs with beat sheets, shot lists, VO scripts, captions, and specs.

Community
Advanced
icoolworldicoolworld

distribution-operator

Generate platform-aware publishing variants for titles, covers, hashtags, and posting times.

Community
Advanced
TheMattBermanTheMattBerman
68

scrollclaw-broll

Generate environment-matched B-roll clips aligned to scripted voice beats.

Community
Intermediate
ChatCut-IncChatCut-Inc
751

create-motion-graphics

Create and edit JSX motion graphic assets in ChatCut projects.

Official
Advanced