How gavinzbf/video_6_codex_fx works
This repository is a Claude Code skill for recreating a product video from a reference video and product image. Its documented workflow analyzes and segments the reference, prepares visual and prompt materials, archives them for review in Lark, then generates and returns the video. The README documents these stages and integrations; source excerpts were not provided, so the graph does not assert implementation-level calls beyond that documented flow.
Main entry points: run_video_breakdown.py Video analysis · segment_router.py Segment routing · storyboard-grid-v1.md Storyboard grid · SKILL.md Claude skill
Components 17
Workflow control
- Claude skill
.agents/skills/product-video-replication/SKILL.md - Video analysis
.agents/skills/product-video-replication/scripts/run_video_breakdown.py - Segment routing
.agents/skills/product-video-replication/scripts/segment_router.py
Video preparation
- Scene anchors
.agents/skills/product-video-replication/prompts/hot-video-replication/scene-anchor-v1.md - Storyboard grid
.agents/skills/product-video-replication/prompts/hot-video-replication/storyboard-grid-v1.md - Image generation
.agents/skills/product-video-replication/scripts/yunwu_image_generate.py - Seedance prompts
.agents/skills/product-video-replication/prompts/hot-video-replication/adapted-replication-plan-v1.md
Review and archive
- Lark archive
.agents/skills/product-video-replication/scripts/push_replication_to_lark.py - User approval
.agents/skills/product-video-replication/references/workflow-gates.md - Lark tables
- Lark video writeback
.agents/skills/product-video-replication/scripts/lark_upload_video.py
External generation
- Dreamina generation
.agents/skills/product-video-replication/scripts/dreamina_generate.py - Yunwu AI
- Kie fallback
Other components
- User
- Reference video
- Product image
Connections 19
- User → Claude skill: requests replication
README.md - Reference video → Video analysis: input video
README.md - Product image → Claude skill: provides product image
README.md - Claude skill → Video analysis: starts analysis
README.md - Video analysis → Segment routing: routes segments
README.md - Segment routing → Scene anchors: multi-part only
README.md - Segment routing → Storyboard grid: prepares visuals
README.md - Storyboard grid → Image generation: generates grid
README.md - Image generation → Yunwu AI: calls image API
README.md - Video analysis → Yunwu AI: uses Gemini
README.md - Video analysis → Kie fallback: fallback analysis
README.md - Storyboard grid → Seedance prompts: informs prompts
README.md - Seedance prompts → Lark archive: archives package
README.md - Video analysis → Lark archive: archives breakdown
README.md - Lark archive → Lark tables: writes records
README.md - User → User approval: reviews archive
README.md - User approval → Dreamina generation: permits generation
README.md - Dreamina generation → Lark video writeback: returns video
README.md - Lark video writeback → Lark tables: writes video
README.md