blender-kiln
Claude Code plugin that turns a text brief or a photo into a production-ready 3D asset through Blender: rebuilt by script, textured from CC0 scans, graded against the real object, optimized to GLB/FBX
写剧本 skill:将小说或相关设定转译为专业商业剧本(国内短剧/出海微短剧双模式)。适配各大 Agent 平台与 AI 视频模型动作解耦。
Translation pending
SkillsPythonGitHub @egguy886Self-installOwner unconfirmed
Paste into Claude or ChatGPT for installation instructions suited to your environment.
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Installable projects can be verified without a deployment domain. Use GitHub sign-in for a personal repository, or the repository verification file above for an organization repository.
Ask about other verification options
The maker's introduction consists of the line above. Read the repository README for more details. Repository README ↗
Loading the full README.
Read the full README ↗English · 简体中文
An agent skill that turns novels, outlines and world-building into production-ready micro-drama screenplays — and the asset bible your AI video pipeline needs. Built for vertical short drama (ReelShort, DramaBox, ShortMax, TikTok, Douyin, Kuaishou), with a dedicated overseas mode: English character names, Chinese action and camera lines for the production team, and natural American English dialogue with Chinese subtext notes.
In AI short-drama production, the screenplay is the foundation of everything downstream. Every character, every outfit, every prop and every location in the script becomes a reference image that has to be generated and locked; every speaking role becomes a voice that has to be cast. A weak script costs you twice: once on screen, and again in wasted generations.
This skill covers the whole front end:
Novel / outline ──► Long-novel compression ──► Episode outline with cliffhangers ──► Scripts ──► Production bible ──► AI video pipeline
(≤60 episodes) (hook rotation, paywall, midpoint) (lint) (characters, wardrobe, (asset library +
props, locations, voices) reference-image prompts)
bible.json. It cross-checks every script, produces a per-episode asset breakdown with generation priority, and exports an asset library plus reference-image prompts for an AI video pipeline.write-screenplay-skill/
├── SKILL.md # Skill definition (agent-facing, Chinese)
├── README.md / README.zh-CN.md # This page, in English and Chinese
├── templates/
│ ├── chinese_script_template.md # Domestic Chinese script template
│ ├── overseas_script_template.md # Overseas script template
│ └── adaptation/ # Long-novel mode: parameters, compression decisions, continuity ledger,
│ # episode outline example, episode template, production bible example
├── references/
│ ├── novel-to-script-engine.md # Novel → script: causal skeleton, externalising interiority
│ ├── overseas-script-protocol.md # Overseas genres and dialogue rules
│ ├── ai-friendly-beat-decoupling.md # Action decoupling for AI video
│ ├── long-novel-adaptation.md # Compression toolbox, episode budget, outline cards
│ ├── episode-hook-system.md # Eight hook types, hook ladder, open-payoff
│ ├── character-and-dialogue.md # Character design, voice sheets, subtext, information tools
│ └── production-bible.md # Characters, wardrobe, props, locations, voices; ID rules; export
└── scripts/ # Zero-dependency Python 3 tools
├── install.py # Link the skill into every supported agent
├── chapter_index.py # Scan chapters → beat-card skeleton (word count, dialogue/text-message share, chapter endings)
├── merge_cards.py # Merge beat cards written volume by volume
├── episode_budget.py # Weighted episode allocation per arc; paywall / midpoint / climax checks
├── outline_check.py # Outline gate: chapter coverage, hook rotation, hook ladder, paywall
├── script_lint.py # Episode gate: format, dialogue budget, runtime estimate, interiority, ending hook
├── bible_check.py # Production bible gate, script cross-check, asset breakdown, pipeline export
└── adapt_common.py
Option 1 — link into every agent at once
git clone https://github.com/egguy886/write-screenplay-skill.git
cd write-screenplay-skill
python scripts/install.py
The installer links the skill into ~/.claude/skills/ (Claude Code), ~/.codex/skills/ (Codex CLI), ~/.gemini/config/skills/ (Antigravity) and ~/.agents/skills/ (other agents).
Option 2 — per project
git clone https://github.com/egguy886/write-screenplay-skill.git .agents/skills/write-screenplay
Trigger it by name or with any of: /写剧本, screenplay, novel-to-script, production bible, character bible, long-novel adaptation.
Examples
Adapt chapter 1 of this novel into three 90-second domestic vertical episodes. Lean on the identity-reversal payoff.
Turn this world-building doc into an overseas micro-drama (ReelShort style): English names, Chinese action, English dialogue with Chinese subtext.
My novel has 138 chapters. Compress it into a 60-episode, 2-minute-per-episode drama. Start with the beat cards and the episode outline.
Build the production bible for the main cast and export it for my AI video pipeline.
| Step | Output | Tool | Gate |
|---|---|---|---|
| L0 Parameters | episode cap, runtime, platform, content level, voice-over policy, red lines | template | user confirms |
| L1 Digest | chapters.json beat cards | chapter_index.py, merge_cards.py | every chapter carded; endings checked against the actual text |
| L2 Compression | subplots, characters, locations, repeated beats, reordering | template | every item has a decision and a reason |
| L2.5 Production bible | bible.json | bible_check.py | no ❌; leads approved by the user |
| L3 Budget | episodes per arc | episode_budget.py | total ≤ cap |
| L4 Outline | episodes.json → readable outline | outline_check.py | no ❌; user reviews every episode |
| L5 Writing | scripts, 5 episodes per batch | script_lint.py, bible_check.py --scripts | no ❌ |
| L6 Ledger | continuity ledger, per-volume read-through | template | ledger matches scripts |
### 场 01. 私人俱乐部宴会厅 · 内 · 夜
**出场人物**:Ivy, Adrian
**造型**:Ivy=C01-GALA, Adrian=C02-SUIT
**核心道具**:旧钥匙扣
△ (中景)大厅门口,Adrian 停下脚步,视线越过人群,落在 Ivy 握杯的右手上。
△ (近景)Adrian 抬起右手,指间挂着一只旧钥匙扣。【停留 1.5 秒】
**Adrian**
(走到她面前,声音很轻)
"Careful. You're shaking."
(中文潜台词:别装了,我认得你。)
**【Episode 001 完 · 集尾钩子:reveal】**
Scene headings use ### 场 NN. and speakers use **Name**, so scripts load directly into downstream dialogue-locking tools. Voice-over (**Name (V.O.)**) and on-screen text (【屏幕 · A → B】) are kept out of the video model and added in post.
Every script is checked before delivery:
bible_check.py --export-h3 writes its asset library format directly.sw-character-conflict, sw-dialogue and sw-series-structure with this skill when you want the underlying theory.MIT. Stars, forks and pull requests are welcome.