qcut-toolkit

# QCut Toolkit

Unified entry point for QCut's six sub-skills. Route tasks to the appropriate sub-skill based on what the user needs.

## Sub-Skills

### 1. native-cli — Project Setup & Native Pipeline Commands
**When:** Setting up a project, cleaning up files, organizing workspace, importing media
**Invoke:** `/native-cli`
**Skill path:** `.claude/skills/native-cli/SKILL.md`

Handles:
- Initializing the standard project layout (`input/*`, `output/*`, `config/`)
- Organizing media by extension with `organize-project`
- Running structure audits with `structure-info`
- Running editor media/timeline/export/diagnostic commands (`editor:*`)
- Running additional native pipeline commands when needed

### 2. ffmpeg-skill — Media Processing
**When:** Converting, compressing, trimming, resizing, extracting audio, adding subtitles, creating GIFs, applying effects
**Invoke:** `/ffmpeg-skill`
**Skill path:** `.claude/skills/qcut-toolkit/ffmpeg-skill/SKILL.md`

Handles:
- Format conversion (MP4, MKV, WebM, MP3, etc.)
- Video compression (`-crf`), resizing (`scale=`), trimming (`-ss`/`-t`)
- Audio extraction, subtitle burn-in, text overlays
- GIF creation, speed changes, merging/concatenation
- Streaming (HLS, DASH, RTMP) and complex filtergraphs

### 3. ai-content-pipeline — AI Content Generation & Analysis
**When:** Generating images/videos/avatars, transcribing audio, analyzing video, running AI pipelines
**Invoke:** `/ai-content-pipeline`
**Skill path:** `.claude/skills/qcut-toolkit/ai-content-pipeline/SKILL.md`

Handles:
- Text-to-image (FLUX, Imagen 4, Nano Banana Pro, GPT Image)
- Image-to-video (Veo 3, Sora 2, Kling, Hailuo)
- Avatar/lipsync generation (OmniHuman, Fabric, Multitalk)
- Speech-to-text transcription with word-level timestamps (Scribe v2)
- Video analysis with Gemini 3 Pro
- YAML pipeline orchestration with parallel execution
- Motion transfer between images and videos

### 4. seedance — Video Prompt Engineering
**When:** Writing video prompts, Seedance/即梦 w...