AI tools, scored for your job
Get paid to train AI
🤝 You're in the Free & Open Source library — community-built tools you can use without paying. A different concept from our scored AI directory.
Home › Free & Open Source › Customer Support › system_prompts_leaks
✦ Free & open sourceCustomer Support

system_prompts_leaks

A public repository of verbatim extracted system prompts and rules from major frontier AI chatbots and tools.

Deconstruct commercial AI prompts and agent rules

Study verbatim system instructions from OpenAI, Anthropic, and Google to improve your own support agents.

Free forever · open source · CC0-1.0
✓ Use today✓ Browser-based✓ Real-time collaboration
9.3
ZEKAI
Essential prompt reference.
Free & open source · Customer Support
Ease of setup10
Usefulness8
Maintenance9.5
Community9.8
★ 65,608
GITHUB
Sep 2026
UPDATED
CC0-1.0
LICENSE
🛠 Don't want to self-host? We can set it up →
⚡ Quick answer

This repository is a curated text collection of captured system prompts from frontier language models and agent tools. If you build AI support agents or conversational flows, you can inspect how major vendors format guardrails, memory, and tool instructions.

Runs on
Cross-platform
Hosting
Desktop
Languages
JavaScript
License
CC0-1.0

What it is

It aggregates raw Markdown files containing system-level instructions captured from ChatGPT, Claude, Gemini, Grok, and coding agents. The files detail tool usage formats, subagent dispatch instructions, persona rules, and injected context structures.

Who it's for

It suits prompt engineers, AI support architects, and product teams designing custom assistant rules or agentic workflows. It is not suitable for users seeking a runnable software application, software library, or automated prompt management tool.

Key features

Vendor Prompt Archives

Examine system instructions across OpenAI, Anthropic, Google, xAI, and Perplexity releases.

Subagent and Skill Layouts

Review how systems like Claude Code and Cowork structure tool definitions and subagent delegation.

Injected Reminders & Memory

Inspect how platforms inject runtime context, memory updates, and safety guardrails mid-session.

Verbatim Markdown Format

Read plain Markdown files directly in GitHub or download them to your local environment.

How to start

Up and running in 3 steps

1
Browse the repository folders corresponding to vendors like Anthropic, OpenAI, or Google.
2
Open specific model or tool Markdown files to inspect raw system instructions and guardrails.
3
Copy and adapt proven context structures into your custom support bot or agent framework.
Real talk

Our honest verdict

What's great

  • Direct insight into production-grade prompt engineering from leading AI vendors.
  • Includes specialized agent prompts covering tools, subagents, and memory systems.
  • Plain Markdown format allows instant search, diffing, and offline reading.
  • Maintained as a public domain dataset without proprietary code restrictions.

Where it falls short

  • Contains raw text files rather than executable software or an active software API.
  • Extracted prompts may change without notice as vendors update their internal model system prompts.
  • Requires manual effort to adapt external prompt structures to your own customer support tech stack.
Cost

Pricing & license

The repository contents are released under the CC0-1.0 public domain dedication, making them free to view, copy, and modify. There are no paid hosting tiers or commercial licenses associated with this text collection.

Paid work

Open remote AI-training roles

880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.

See all open roles →
Compare

Alternatives

Get more

Use it, or have us set it up

📄 Free PDF: system_prompts_leaks for Customer Support
A plain-language guide to getting your first result. Sent to your inbox.
🔔 64 people want a full video tutorial — add your vote
🛠 Don't want to set it up yourself?
We'll install, host & configure system_prompts_leaks for you — and train your team.
Get it done for you ↗
Questions

FAQ

How can this help build better customer support bots?

You can analyze how top AI companies set system guardrails, handle edge cases, format function calls, and enforce brand tone.

Is there any executable code in this repository?

No, it consists primarily of Markdown files containing extracted prompt text and configuration instructions.

Radar

Latest AI news

✉️ The weekly free toolkit
The best new free & open-source tools for Customer Support — one short email a week, in your language.
Free · 1 email/week · unsubscribe anytime
Get paid to train AI →

See Zekai first in Google