main/CONTRIBUTING.md

Contributing to CL4R1T4S

Thank you for helping make AI system prompts more transparent! Every contribution helps the community understand how AI systems are instructed behind the scenes.

How to Submit a Prompt

1. Fork this repository
2. Add your prompt file to the appropriate company directory (see Directory Structure)
3. Open a pull request with a clear description

File Requirements

Every submitted prompt file should include:

• ✅ Model name and version (in the filename and metadata header)
• 🗓 Date of extraction in ISO 8601 format (YYYY-MM-DD), if known
• 🧾 Context or notes (optional but helpful)

Metadata Header

Please add a metadata header at the top of every prompt file:

---
model: "Claude 4.1"
company: "Anthropic"
date_extracted: "2025-08-05"
extraction_method: "Direct prompt extraction"
contributor: "your-github-username"
notes: "Extracted via conversation prompt leak"
---

If you don't know a field, omit it or write unknown.

Naming Conventions

Filenames

• Use underscores (_) as word separators: Claude_Sonnet_3.5.md
• Include the extraction date in ISO 8601 format when known: ChatGPT_4o_2025-04-25.txt
• Use .md or .txt file extensions — never omit the extension
• Format: ModelName_Version_YYYY-MM-DD.md

Dates

• Always use ISO 8601: YYYY-MM-DD (e.g., 2025-04-25)
• Never use Month-DD-YYYY, MM-DD-YY, or other formats

Directory Names

• Use ALL CAPS for company/product directory names: OPENAI, ANTHROPIC, XAI
• Use hyphens instead of spaces: BURGESS-PRINCIPLE, VERCEL-V0

Directory Structure

Each top-level directory corresponds to a company or product:

ANTHROPIC/       — Anthropic (Claude models)
BOLT/            — Bolt
BRAVE/           — Brave (Leo)
BURGESS-PRINCIPLE/ — The Burgess Principle framework (not a prompt)
CLINE/           — Cline
CLUELY/          — Cluely
CURSOR/          — Cursor
DEVIN/           — Devin
DIA/             — Dia
FACTORY/         — Factory (DROID)
GOOGLE/          — Google (Gemini models)
HUME/            — Hume
LOVABLE/         — Lovable
MANUS/           — Manus
META/            — Meta (Llama models)
MINIMAX/         — MiniMax
MISTRAL/         — Mistral (Le Chat)
MOONSHOT/        — Moonshot (Kimi models)
MULTION/         — MultiOn
OPENAI/          — OpenAI (ChatGPT, GPT, Codex, Atlas)
PERPLEXITY/      — Perplexity
REPLIT/          — Replit
SAMEDEV/         — SameDev
VERCEL-V0/       — Vercel v0
WINDSURF/        — Windsurf
XAI/             — xAI (Grok models)

If your prompt is from a company not listed here, create a new ALL-CAPS directory for it.

What Counts as a Contribution?

• System prompts — the hidden instructions that shape model behavior
• Tool definitions — function/tool schemas given to models
• Guidelines or policies — internal moderation or behavior guidelines
• Personality/persona configs — instructions that define how a model presents itself

Community

• X (Twitter): @elder_plinius
• Discord: Reach out to @elder_plinius

License

By contributing, you agree that your contributions will be licensed under the GNU AGPL v3 license that covers this repository. The Burgess Principle directory is separately licensed under MIT (see LICENSE-BURGESS).