# poetry-cinema-page

poetry-cinema-page turns a classical Chinese poem into a scroll-driven page with a matched image set, generated narration and a line-by-line reading.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf (Devin Desktop), Claude.ai · ★ 297 repo stars
- Author: wwwzhouhui
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/wwwzhouhui/skills_collection/tree/HEAD/poetry-cinema-page/SKILL.md
- Updated: 2026-10-29
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/skills/poetry-cinema-page

## What it does

The skill takes a poem, ci, qu or prose-poem and lays it out as film-like scenes on a single page. It inspects the repo first (framework, build scripts, existing routes) and stays on the current stack. The text is then split into 5–8 visual beats grouped by imagery and emotional turns, not by equal line counts.

Images come from the bundled CLI `scripts/ark_image.py`, which talks to two providers: Volcengine Ark with Doubao Seedream 5.0 and a third-party GPT-Image gateway. The tier decides the provider, so both can appear in one poem. A hero frame is generated first with text-to-image; every beat after it goes through image-to-image with the hero as reference, and that is what holds period and palette together. Run `python scripts/ark_image.py probe` before spending anything — it resolves keys and model ids without billing an image.

Narration is produced, not supplied: `edge-tts` runs locally with no key, or Doubao seed-tts uses the same Ark key. A Remotion render can turn finished scenes and subtitles into a voiceover video on the narration timeline. One trap on the gateway: reference images work only as multipart `POST /images/edits`. Point the call at `/images/generations` and the reference fields are dropped silently, so you get a 200 and a fresh picture with the chain gone.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add wwwzhouhui/skills_collection --skill poetry-cinema-page -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/poetry-cinema-page/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/poetry-cinema-page/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add wwwzhouhui/skills_collection --skill poetry-cinema-page -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/poetry-cinema-page/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add wwwzhouhui/skills_collection --skill poetry-cinema-page -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/poetry-cinema-page/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install wwwzhouhui/skills_collection poetry-cinema-page
```

2. **Manually: put the files in place**
- File: `.github/skills/poetry-cinema-page/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add wwwzhouhui/skills_collection --skill poetry-cinema-page -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/wwwzhouhui/skills_collection --path poetry-cinema-page
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/poetry-cinema-page/SKILL.md` — The legacy .windsurf/skills/ path still works

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/wwwzhouhui/skills_collection/tree/HEAD/poetry-cinema-page/SKILL.md

## When to use it
- build a scroll-driven long-form page from a poem
- keep one visual world across every generated image
- get a spoken narration without recording a voice
- check keys and model ids before paying for images
- cut finished scenes into a subtitled narrated video

## Example prompts
- Make a cinematic page for this poem: 5–8 beats, and run `python scripts/ark_image.py probe` first.
- Generate the narration with `edge-tts` and cut the finished scenes into a Remotion video with subtitles.
- Regenerate scene 3 so the character still matches `hero.jpg`.

## Similar
- [Poetry Packaging](https://affpapa.org/en/nevibecode/skills/poetry-packaging.md) — skill · 132 installs
- [poetry](https://affpapa.org/en/nevibecode/rules/poetry.md) — rule · ★ 3589 repo stars
- [page-mascot](https://affpapa.org/en/nevibecode/skills/page-mascot.md) — skill · 1500 installs
- [Soundside-ai](https://affpapa.org/en/nevibecode/mcp/soundside-ai.md) — MCP server · 815 installs
- [scroll-world](https://affpapa.org/en/nevibecode/skills/scroll-world.md) — skill · 5400 installs
- [video-toolkit](https://affpapa.org/en/nevibecode/skills/video-toolkit.md) — skill · 1200 installs

From the same set: [wwwzhouhui/skills_collection](https://affpapa.org/en/nevibecode/by/wwwzhouhui/skills_collection.md) — 7 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
