ai-shortfilm-prompts skill
A Claude Code skill that turns an idea into a five-stage prompt for Sora, Kling, Veo and Seedance, with templates for 21 genres
- 499 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill consists of instructions and templates and neither runs code nor calls the network
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/ai-shortfilm-promptsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill consists of instructions and templates and neither runs code nor calls the network
This is third-party code. Review the repository files before installing.
What it does
ai-shortfilm-prompts is a repository with the shortfilm-prompt skill for Claude Code. The skill turns an idea into a ready prompt for the Sora, Kling, Veo, Seedance, Jimeng and Xiaoyunque video models. The prompt follows five stages: a theme in 3 to 6 tags, character and scene, atmosphere and quality, camera rules, and a storyboard by second or by shot. When the request is thin, the skill asks two or three questions: video type and duration, subject and scene, visual style. It then loads the template for the matching genre out of 21 (a 15-second transformation, a car commercial, found footage, a tearful pet story and others), writes the prompt and checks it against a ten-item list. At the end it explains two or three of its choices and adds a line of advice for the chosen model: Seedance, for example, blocks the names of other people's franchises, while Sora does better with short prompts. Besides the skill, the repository holds the full original prompts of the author of the short film Zombie Scavenger, a one-page cheat sheet, a gallery of common failures with fixes, and a set of test cases in the evals folder.
Who it is for. For makers of short AI videos, advertisers and content creators who write prompts for video models and want a ready structure instead of trial and error.
Good fit when
- You need a 15-second prompt for a hero or robot transformation with a scene and color palette
- You have an idea but no structure: the skill splits it into theme, scene, atmosphere, camera and storyboard
- A Seedance prompt gets rejected over a franchise name and you need wording without it
- You want a multi-shot scene, a tearful pet story or a product commercial drafted from a genre template
- You need a list of typical generation failures and their fixes, which the cases.md gallery provides
Not a fit when
- You need the video itself: the skill writes only the prompt text, generation is done by the video model
- You need editing or work on a finished clip: the catalog has ffmpeg-based video skills for that
- You need a prompt for images or text: the templates target video models
- You need quality guarantees: results depend on the model and the number of attempts, the author offers a method, not a promise
Example request
/ai-shortfilm-prompts:shortfilm-prompt Write a 15-second prompt: a robot transformation in a green palette, an energy core in the belt buckle, a post-apocalyptic jungle backgroundLimitations
The skill works with text only: it creates no video and does not connect to video models. The models are paid for and connected separately, and OpenAI services are unavailable from Russia. In the compatibility table the author marks the durations for Seedance and Hailuo as approximate, and some service links in it are invitations with bonuses. A good prompt from this method still takes several attempts, and generation cost depends on the model you pick. The NOTICE file sets separate terms for Mx-Shell's original prompts, so check them before commercial use.
How to disable. If you installed through the marketplace, run /plugin and disable or remove ai-shortfilm-prompts. If you copied it by hand, delete the ~/.claude/skills/shortfilm-prompt folder.
FAQ
How do I install the video prompt skill in Claude Code?
Add the marketplace with /plugin marketplace add jnMetaCode/ai-shortfilm-prompts and install the plugin: /plugin install ai-shortfilm-prompts@ai-shortfilm-prompts. You call the skill as /ai-shortfilm-prompts:shortfilm-prompt with a description of the clip. There is also a manual route: copy the skills/shortfilm-prompt folder into ~/.claude/skills.
Why do the prompts contain no franchise or character names?
Seedance 2.0 blocks the names of other people's franchises. The skill writes a made-up trigger phrase instead, such as a whispered syllable of its own invention. Kling is more permissive about such names, and the skill notes that in its per-model advice line.
Can I use it without Claude Code?
Yes. The author published an online builder: pick one of 21 genres, fill in a few fields and get a full multi-shot prompt with a negative prompt. It runs in the browser, and the templates also live in the templates folder.
Which video models does the five-stage structure work with?
The compatibility table lists Seedance 2.0, Veo 3 and 3.1, Kling 2.x and 3.0, Hailuo, Wan 2.x, Runway Gen-4, Pika and Sora 2. For each it gives the single-shot limit, negative prompt support and the franchise filter. Runway, per the table, does not support negative prompts, and the skill asks you to describe only what should appear.
What else is in the repository besides the skill?
A templates folder with 21 genre templates in English and Chinese, a prompts folder with Mx-Shell's original prompts, the cheatsheet.md cheat sheet, the cases.md failure gallery and methodology.md. The evals folder holds test cases and the run_evals.py script.
Discussion
No questions or reviews yet.
Related
SkillsSeven Claude Code skills and a prompt library for the Seedance, Kling and HappyHorse video models: prompts, storyboards, debugging
A skill that writes prompts for video, image and music generation on 14 platforms and can submit them in the browser
A skill that teaches an agent to write prompts for ByteDance's Jimeng Seedance 2.0 video model: @ references, camera language, timing, templates
A set of skills for Claude Code and Codex that assemble detailed image prompts from an element library and composition schemes
Five Claude Code skills that produce faceless vertical YouTube Shorts from Remotion code, video models or paper collage
A Claude Code skill that turns one sentence into a vertical AI micro-drama, from script and storyboard to video
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| ai-shortfilm-prompts (this entry) | no | 499 | |
| Awesome Seedance | no | 2,344 | a collection of tested ready prompts for Seedance, without a skill that writes a prompt for your task |
| Seedance 2.0 Prompt Writing Skills | no | 4,376 | a skill for Seedance and Jimeng only, while here the structure is shared across eight models |
| OnlyShot | no | 341 | a Claude Code skill that turns one sentence into a vertical AI |
README in short
The README opens with links to the online prompt library and builder, then tells the story of the short film Zombie Scavenger, from which the author derived the method. It continues with a one-line prompt to try, a weak versus strong prompt comparison, three rules (name real cameras and lenses, describe imperfections, leave the ending empty) and the five-stage scheme. It closes with four install routes, a table of compatible video models and links to the author's related projects.
README badge
Original SKILL.md
--- name: shortfilm-prompt description: Generate cinematic AI shortfilm prompts (works with Seedance 2.0, Xiaoyunque, Sora, Kling, Jimeng, Veo) using the 5-stage structure from Mx-Shell's Zombie Scavenger. Trigger when the user wants transformation sequences, multi-shot narrative shorts, weapon-charge/combat segments, emotional family/pet/farewell narratives (催泪/亲情/萌宠/离别), or any cinematic video prompt. --- # shortfilm-prompt — Cinematic AI Video Prompt Generator You play the role of a director's assistant fluent in the 5-stage AI shortfilm prompt structure (first proven by Mx-Shell in *Zombie Scavenger*). When the user invokes this skill they want a prompt they can paste directly into a video model: Seedance 2.0 / Xiaoyunque / Sora / Kling / Jimeng / Veo. **Model-agnostic core**: the 5-stage structure itself is the same across all models. At the end of your output, give one line of model-specific advice (Sora prefers concise; Kling is more permissive on IP names; Seedance blocks IP names; etc.). ## Workflow (execute in order) ### Step 1 — Did the user already specify enough? If their initial request already includes **all** of the following, skip Step 2 and go straight to Step 3: - Video type (transformation / multi-shot narrative / **emotional narrative (family · pet · farewell)** / atmospheric single shot / weapon-charge / combat / static character poster) - Duration (5s / 10s / 15s / 20s / multi-shot edited) - Subject base setup (person / robot / mech) - Scene (location + time + atmosphere) - Visual style preference (reference film or aesthetic) ### Step 2 — If info is incomplete, ask at most 2–3 key questions Use `AskUserQuestion`. Priority order: 1. **Video type + duration** (decides which template branch) 2. **Subject + scene** (decides content) 3. **Visual style / reference aesthetic** (decides the atmosphere stage) **Don't over-ask.** Mx-Shell himself worked iteratively, making it up as he went. Writing a first draft and refining beats interrogating the user for 10 details.