Prompts library

Prompts you can drop into Cursor / Claude Code / ChatGPT / Gemini to drive cutcli.

Core prompts

Type	Purpose	File
System	"You are a cutcli expert" — fits ChatGPT custom instructions	prompts/system/cutcli-expert.md
Cursor	One-line generation of a product promo inside Cursor	prompts/cursor/make-promo-video.md
Claude	Auto-storyboarding then cutcli command emission	prompts/claude/auto-storyboard.md

Usage

Cursor / Claude Code

No copy-paste required, just run:

cutcli setup cursor   # or claude / openclaw / all

Details: AI tools integration.

ChatGPT / Gemini / generic LLM

Copy the full content of prompts/system/cutcli-expert.md into the model's "system prompt" or "custom instructions". You can then ask freely: "Make me a 5-second caption with fade-in" — the model will produce a runnable cutcli command.

Inside an automation script

Wire it into an OpenAI / Anthropic API call as the system message:

import anthropic, pathlib

system_prompt = pathlib.Path("prompts/system/cutcli-expert.md").read_text()
client = anthropic.Anthropic()
resp = client.messages.create(
    model="claude-sonnet-4",
    system=system_prompt,
    messages=[{"role": "user", "content": "Build me a 5-second Hello World caption."}],
)
print(resp.content)

Contributing prompts

Add new prompts under prompts/<tool>/<scenario>.md with the following frontmatter:

---
title: One-line title
author: your-handle
target: cursor | claude | chatgpt | gemini | openai-api
tags: [marketing, vlog, ...]
license: MIT
---

Open a PR — CI will validate the format.