# asgeirtj/system_prompts_leaks

Extracted system prompts from Anthropic - Claude Fable 5, Opus 5, Claude Design, Claude Code. OpenAI - ChatGPT GPT-5.6-Sol, Codex. Google - Gemini 3.5 Flash, 3.1 Pro, Antigravity. xAI - Grok, Cursor, Copilot, VS Code, Perplexity, and more. Updated regularly.

Repository: https://github.com/asgeirtj/system_prompts_leaks
Canonical: https://ross.abutalabs.com/products/system_prompts_leaks
Language: JavaScript
License: CC0-1.0
License Family: permissive
Topics: ai, anthropic, chatbot, chatgpt, claude, claude-code, codex, gemini, generative-ai, google, llm, openai, prompt-engineering, ai-agents, system-prompts, cursor, grok, ai-prompts, prompt, system-prompt
Last push: 2026-08-25T07:01:45+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 35, longevity 34
- inputs: {"age_days": 487, "days_push": 8, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 63624, forks 10419 (observed 2026-08-28T04:12:19.168257+00:00)

## What it is
A curated, regularly updated collection of verbatim leaked system prompts from major AI products including ChatGPT, Claude, Gemini, Grok, Cursor, Copilot, and Perplexity. It serves as a reference dataset for studying how commercial LLM products instruct and constrain their underlying models.

## Use cases
- study how ChatGPT or Claude system prompts are structured
- research prompt engineering patterns used by AI vendors
- compare system prompts across models like Gemini, Grok, and Claude
- build a dashboard or analysis of AI chatbot hidden instructions
- learn how coding agents like Claude Code or Codex are prompted
- find examples of tool definitions and skills in production prompts

## When to choose
- you want real, verbatim system prompts from major AI products
- you're researching LLM behavior, safety rules, or prompt design
- you need a regularly updated corpus of production prompts
- you're writing about or teaching prompt engineering with real examples

## When to avoid
- you need prompts guaranteed to be current or official — leaks may be outdated or altered
- you want a prompt library to copy into your own app rather than study
- you require a structured, machine-readable dataset with schemas
- your use case depends on legally sanctioned data sources

## Facets
- artifact type: dataset
- maturity: active
- function: prompt-engineering, llm-inference, developer-tools
- domain: large-language-models, artificial-intelligence, chatbots, tutorials
- platform: cross-platform
- tags: system-prompts, leaked-prompts, prompt-collection, llm-behavior, reference-dataset, chatgpt, claude, gemini, grok, ai-agents

## Member repositories
- asgeirtj/system_prompts_leaks (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:12:19.168257+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:17:26.644223+00:00, confidence not recorded.
  - readme: https://github.com/asgeirtj/system_prompts_leaks (fetched 2026-08-28T04:12:19.168257+00:00, sha 7efc60f18a8c)
- Data as of 2026-08-30T08:39:29.467469+00:00.