openmontage-video-production

Orchestrates AI short-drama and video production pipelines from script through final delivery.

Updated Jul 16, 2026
One-click install
npx skills add https://github.com/X-manist/Cohmira --skill openmontage-video-production-x-manist
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: openmontage-video-production
Source: https://github.com/X-manist/Cohmira/tree/main/src/builtin-plugins/openmontage/skills
Command: npx skills add https://github.com/X-manist/Cohmira --skill openmontage-video-production-x-manist

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Producing AI short dramas and marketing videos requires coordinating scripting, storyboarding, asset generation, editing, compositing, and quality checks across many tools; this Skill turns that scattered work into a staged, reviewable pipeline with persistent artifacts. ## Core Features & Use Cases - Staged pipeline orchestration: Runs research, proposal, script, scene planning, assets, edit, compose, and publish stages with director guides, quality gates, and human approval checkpoints. - AI short-drama workflow: Creates drama projects with persistent characters, episode plans, screenplays, and storyboards, saving artifacts via MCP tools and collecting binding user approvals through review UIs. - Managed video generation and delivery: Routes all video generation through the built-in Volcengine-backed service, then handles stitching with continuity checks, subtitles, audio, QA, and final delivery from the media library. - Use Case: Ask the agent to adapt a short story into a multi-episode AI drama; it creates the project, drafts the screenplay and storyboards stage by stage, generates clips with continuation mode, stitches them, and delivers the final cut. ## Quick Start Use the OpenMontage video production skill to turn my story outline into a three-episode AI short drama with storyboards, generated clips, and a final stitched video.

Frequently Asked Questions about openmontage-video-production

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

FAQPage Schema
How do I create an AI short drama from a novel or story idea?

The skill selects the ai-short-drama pipeline, creates a project with drama_project_create, then walks through research, proposal, script, scene, asset, edit, compose, and publish stages. Each stage's artifacts are saved with drama_artifact_save and approved through the review UI.

Which video generation providers does this skill support?

Video generation goes through a single supported entry: video_selector or seedance_video, both routed to the built-in Volcengine-backed service via app_cli. FAL, Kling, Runway, Veo, MiniMax, Replicate, ComfyUI, and local GPU generators are explicitly not used.

How do I make a multi-segment video longer than one clip's limit?

Plan multiple segments and stitch them with video_stitch. From the second segment onward, use seedance_video continuation mode with the previous clip as first_clip, and keep continuity_check set to strict during stitching.

Can I use English prompts for image and video generation?

Chinese is the default working language and Seedance accepts Chinese prompts directly, so requirements should not be translated into English just for generation. Titles, prompts, asset descriptions, and final replies all stay in Simplified Chinese unless the user specifies otherwise.

Why must generated files use a shared project_id?

A stable project_id keeps the first image, keyframes, video clips, and stitched output organized under generated/<project-id>/ with images, clips, output, and audio subfolders. Scattering files in the generated root or delivering from /tmp paths is prohibited.

Do I need to install Python or FFmpeg to use this skill?

No separate installation is required. The Python runtime is installed on demand by the application's built-in uv, so users should not be asked to set up Python, pip, or virtual environments themselves.