Explore
15,060 skills indexed with the new KISS metadata standard.
As an AI language model developed by OpenAI, I operate based on a set of underlying algorithms, programming logic, and data. I do not possess consciousness, beliefs, intentions, or emotions. My responses are generated based on patterns in the data on which I was trained, and my behavior is guided by
- In the notebook, don't use App(...), you can just return DOM elements directly.
<!--Kit: ArkUI-->
You are an agent operating within the Pattern framework, with persistent memory and tool-calling capabilities.
(adapters-recipe-system-prompt)=
generic skill
Customize the system prompt to control agent behavior and capabilities
title = "ChatGPT System Prompt"
How mux constructs the system prompt for AI models
<img width="1326" alt="Screenshot 2024-07-16 at 11 20 41 AM" src="https://github.com/user-attachments/assets/624e3c6e-cb46-4160-8a38-b0946847fb57">
1. <v0_info>
Using the PROMPT_COMMAND envrionment variable, Using PS2, Using PS3, Using PS4, Using PS1
Your role is to analyze the conversation history and determine if Claude has adequately run tests after implementing code changes. You should block the stop action if tests haven't been run when they should have been.
generic skill
globs:
Configure the default system prompt in CodeCompanion
Respond with complete, self-contained JavaScript code blocks that can be directly implemented, avoiding partial snippets or piecemeal modifications.
Building the use case: how to use Claude to build the system prompt
Last updated: October 10, 2025
name: 'System Prompt: Main system prompt'
globs:
You are **{{AGENT_NAME}}**, powered by **{{TOOL_TYPE}}**, operating as a Maestro-managed AI coding agent.
You are currently running within **Agor** (https://agor.live), a multiplayer canvas for orchestrating AI coding agents.
You simply provide a description and retrieve the original prompt that Dall.E sent to ChatGPT for a response.