skills.rest
HKUDS
HKUDSOFFICIAL
Explore all skills by this author →
WebsiteGitHub

cli-hub-matrix-video-creation

Orchestrates multi-provider video production capabilities from storyboard planning through final render review.

48.8k|4.5k|Updated Mar 8, 2026
Search Tags:#captions#video production#sound design#FFmpeg#nle#Storyboard#provider matrix
One-click install
npx skills add https://github.com/HKUDS/CLI-Anything --skill cli-hub-matrix-video-creation
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: cli-hub-matrix-video-creation
Source: https://github.com/HKUDS/CLI-Anything/tree/main/cli-hub-matrix/video-creation
Command: npx skills add https://github.com/HKUDS/CLI-Anything --skill cli-hub-matrix-video-creation

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Producing polished videos with AI agents requires coordinating many tools—downloaders, generators, NLEs, caption engines, and audio tools—without a fixed pipeline. This Skill provides a capability-based matrix that lets agents pick the right provider per task instead of marching through rigid stages.

Core Features & Use Cases

  • Capability-based provider selection: Choose providers (CLI harnesses, public CLIs, Python libs, native binaries, cloud APIs) per capability such as storyboard planning, video/music search and download, capture, generation, transcription, caption design, sound design, assembly, and packaging.
  • Preflight and install commands: Run cli-hub matrix preflight video-creation to detect available providers and cli-hub matrix install video-creation to install registered harness CLIs.
  • Quality gates and doctor tooling: Enforce creative direction documents, source triage manifests, caption QC, and render investigation via the bundled video_doctor.py script and reference guides.
  • Use Case: An agent asked to produce a 60-second sports hype video plans a storyboard, downloads licensed footage with yt-dlp, designs ASS captions, builds a sound design with risers and hits, assembles in Kdenlive, and validates the final MP4 with the render doctor.

Quick Start

Ask the agent to create a video from your concept and have it run cli-hub matrix preflight video-creation to select providers for each production capability.

Frequently Asked Questions about cli-hub-matrix-video-creation

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

FAQPage Schema
How do I create a video with AI agents using multiple tools?▼

Use the capability matrix to pick a provider per stage: plan with agent-native storyboarding, download footage with yt-dlp, transcribe with Whisper, design captions in ASS, assemble with Kdenlive or ffmpeg, then validate with video_doctor.py. Run cli-hub matrix preflight video-creation first to see what is available.

What tools can I use to download YouTube or Bilibili videos for editing?▼

yt-dlp is the primary provider for YouTube and Bilibili downloads, with you-get, lux, and BBDown as alternatives. BBDown is preferred when Bilibili metadata, subtitles, danmaku, or member streams matter. Always record license and provenance in sources.json.

Does the video matrix support offline video editing?▼

Yes, many providers are offline-capable including ffmpeg, MoviePy, Whisper, PySceneDetect, Kdenlive, and Shotcut. Run cli-hub matrix preflight video-creation --offline to filter the matrix to providers that work without network access.

How do I add styled subtitles to a video with ffmpeg?▼

Author captions as an ASS file with named styles, set PlayResX/PlayResY to the final render resolution, then burn them with ffmpeg -vf subtitles=captions.ass. The captions reference guide defines genre presets, safe zones, and QC steps before delivery.

When should I use an NLE like Kdenlive versus plain ffmpeg for assembly?▼

Choose Kdenlive or Shotcut when the timeline has many clips, tracks, transitions, or the user wants a reusable project file. Use ffmpeg or MoviePy for deterministic tasks like concat, trim, burn-in, and muxing. The matrix documents a provider-boundary pattern combining both.

Why does my downloaded Bilibili audio sound silent after conversion?▼

yt-dlp -x --audio-format mp3 can produce a valid-looking but nearly silent MP3 from Bilibili sources. Download the raw m4a instead, convert with ffmpeg, then verify with volumedetect and reject files with mean volume below roughly -40dB.

Related Skills

Explore More Design & Creative Skills→
DrinkBooozDrinkBoooz

video-editing

Creates AI-assisted video editing workflows from raw footage to finished clips.

Community
Advanced
adcontextprotocoladcontextprotocol
18

build-creative-agent

Manage creative assets and build VAST or display serving tags.

Official
Advanced
coltonbattscoltonbatts
22

vcr-debugger

Diagnose VCR manifest failures and inspect frame-level rendering state.

Community
Intermediate
CellCogCellCog
6

video-cog

Orchestrate multi-model tasks to produce videos from a single prompt.

Official
Advanced