SuperPrompt prompt
An experimental set of structured prompts for Claude: XML markup, the think tag and several variants in Lisp and TM formats. No license
- 6,424 ★ on the repository
- Last commit
Minimal access
Why this level
- The project is experimental and its effectiveness is not proven
- No license is stated in the repository
Install
- Works in
Manual install
git clone https://github.com/NeoVertex1/SuperPrompt.gitThere is no installation: read the prompt files. Take them into working projects only after testing.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The project is experimental and its effectiveness is not proven
- No license is stated in the repository
This is third-party code. Review the repository files before installing.
What it does
The repository holds experimental prompts for Claude: the main README, the lisp_prompt.md and tm_prompt.md variants, a GPT version in prompt_for_gpt.md, an alternative variant, the flowchart.md diagram and a research folder. The prompts are built on XML markup and annotations, with a <think> tag inside for reasoning. The author himself calls the project an attempt to engineer prompts that help understand AI agents and admits it is in perpetual beta.
Who it is for. For prompt engineers who study non-standard system prompt formats and want to look at one of the popular experiments.
Good fit when
- You want to see how a prompt is built on XML markup with a reasoning block
- You need examples of experimental variants to compare with ordinary prompts
- You need a starting point for your own experiments with prompt format
Not a fit when
- You need a proven working prompt for a product: the project is experimental
- You need a clear license: none is stated in the repository
Example request
Read lisp_prompt.md and list how its structure differs from an ordinary system promptLimitations
The project is experimental, there is no rigorous effectiveness evaluation, and we did not check the author's claims about understanding AI agents. No license is stated in the repository, so rights to the code are not granted by default. The author calls the project a soft jailbreak, so do not use it to get around model rules. The repository was last updated on 26.04.2026.
How to disable. Simply do not use the prompt in your projects. A repository clone is removed by deleting the folder.
FAQ
What are these prompts?
Experimental structured prompts for Claude on XML markup with the think tag. There are Lisp and TM variants and a GPT version.
Is there evidence of effectiveness?
No, the project is experimental and called perpetual beta by the author himself.
What is the license?
No license is stated in the repository.
Can I use it in a product?
Only after your own testing and given the absence of a license.
Discussion
No questions or reviews yet.
Related
PromptsClaude Code Prompts (RepoWise)
claude-code-prompts
Independently written prompt templates for AI coding agents: system prompts, tool prompts, delegation, memory and multi-agent coordination
Awesome Claude Prompts
awesome-claude-prompts
A prompt collection for Claude from the author of LangGPT: business ideas, texts, code, plus a claudecode folder. Part of the examples date from 2023
claude-code-system-prompts
Claude Code System Prompts
A reference of Claude Code system prompts: parts of the system prompt, tool descriptions and built-in agent prompts
Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent
A structured prompt system: templates with role, profile, skills, rules and workflow for repeatable model requests
A skill that writes precise prompts tailored to a specific tool: Cursor, Claude Code, Midjourney, Stable Diffusion and more
README in short
The repository holds a README, prompt variants in Lisp and TM, a GPT version, a diagram and a research folder. The author calls the project an attempt to engineer prompts and keeps it in perpetual beta.