video-production

这段内容是审核时保存的快照。外部来源才是完整且最新的版本。

---
name: video-production
description: >
  Plan and route programmable or automated video production across code-first,
  template-first, and hybrid content pipelines. Use when the user needs repeatable
  video generation, branded short-form content, personalized videos, social clip
  batches, captioned/localized variants, video APIs, or video creation from data
  and templates — even if they only say video production. Also the direct owner of
  explicit Remotion requests (React video compositions, scenes, render workers) and
  of the retired `remotion-video-production` name. Triggers on: Remotion, Remotion
  render pipeline, React video composition, programmatic video, automated video
  creation, video API, personalized video, batch-create shorts, render videos from
  code, repurpose content into clips.
allowed-tools: Write Read WebSearch WebFetch Task
compatibility: >
  Canonical programmable-video / automated-video skill for the repo and the single
  entry point for Remotion-named requests; the former `remotion-video-production`
  alias was merged here and is listed in `skills.json` retired_skills.
metadata:
  tags: video, automation, remotion, short-form, content-ops, templates, react
  platforms: Claude, ChatGPT, Gemini, Codex
  version: "2.0"
---

# Video Production

Use this skill as the **canonical programmable-video and automated-video production anchor** for the repository.

The job is not to dump a generic storyboard or act like a manual editor tutorial. The job is to:
1. normalize the production request,
2. choose the best production mode,
3. return one implementation-ready packet,
4. leave explicit asset, QA, and handoff guidance.

Read [references/production-modes.md](references/production-modes.md), [references/asset-and-qa-checklist.md](references/asset-and-qa-checklist.md), and [references/handoff-boundaries.md](references/handoff-boundaries.md) before routing broad requests.

## When to use this skill
- The user wants automated, repeatable, or template-driven video production
- The request involves short-form content ops, campaign variants, personalized videos, localized versions, or batched social assets
- The user mentions Remotion, rendering from code, video APIs, templates, or connecting app/product data into video output
- The workflow needs a production plan that can span generation, asset prep, QA, and a publishing handoff

## When not to use this skill
- The job is purely a one-off manual edit in Premiere / After Effects / CapCut with no automation or repeatability goal
- The task is only final creative polish, color, taste-based pacing, or bespoke motion-design critique
- The user needs only transcript cleanup, clip selection, or direct publishing with no template/render layer
- The request is really an ad-strategy, content-strategy, or product-launch brief rather than a video-production workflow

## Supported production modes
Use these as routing targets inside the skill:

### 1) Code-first programmable video
- Best when the user explicitly wants Remotion, React-based composition, dynamic data injection, or custom product-integrated rendering
- Default stack: `Remotion`

### 2) Template/API automation
- Best when the user needs speed, bulk generation, localization, or personalized variants without owning a full rendering codebase
- Typical comparators: Shotstack, Creatomate, Bannerbear

### 3) Hybrid clip / repurposing pipeline
- Best when the source material is long-form audio/video and the main problem is extracting, captioning, resizing, and packaging clips at volume
- Common workflow shape: transcript or clip-selection tool + template/render layer + manual QA

### 4) Manual-finish hybrid
- Best when automated generation exists but editorial polish, captions, timing, or final approvals still need a human pass
- Use this when the right answer is not “fully automate everything” but “automate the repeatable 80%, then define the last-mile edit pass"

## Instructions

### Step 1: Normalize the produc