# jujumilk3/leaked-system-prompts

Collection of leaked system prompts

Repository: https://github.com/jujumilk3/leaked-system-prompts
Canonical: https://ross.abutalabs.com/products/leaked-system-prompts
License Family: other
Topics: prompt, llm, ai, document
Last push: 2026-06-20T01:13:25+00:00

## Health v2 (maintenance only)
Score: 69/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 88, release rhythm 35, longevity 86
- inputs: {"age_days": 1206, "days_push": 75, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 14912, forks 2112 (observed 2026-08-28T04:11:08.926296+00:00)

## What it is
A curated collection of leaked system prompts from widely used LLM-based services, with community contributions verified via pull requests and issues. It serves as a reference dataset for studying how commercial AI products instruct their models.

## Use cases
- study how commercial LLM services write system prompts
- research prompt engineering patterns used by AI products
- find real-world examples of system prompt design
- cite examples of leaked prompts in academic papers
- compare system prompts across different AI assistants
- learn prompt engineering from production systems

## When to choose
- you need real-world system prompt examples for research or study
- you want to analyze how popular LLM products instruct their models
- you need citable sources of system prompts for papers

## When to avoid
- you need a licensed, legally cleared dataset for commercial use
- you want an official or authorized prompt library
- you need executable tooling rather than reference documents

## Facets
- artifact type: dataset
- maturity: active
- function: prompt-engineering, llm-inference, documentation
- domain: large-language-models, artificial-intelligence, security, tutorials
- platform: -
- tags: system-prompts, leaked-prompts, llm-services, research-reference, prompt-collection, web-server

## Member repositories
- jujumilk3/leaked-system-prompts (main) score 69

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:08.926296+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 2026-08-29T17:06:55.181968+00:00, confidence not recorded.
  - readme: https://github.com/jujumilk3/leaked-system-prompts (fetched 2026-08-28T04:11:08.926296+00:00, sha 6e8c8b03ce94)
- Data as of 2026-08-30T08:39:29.467469+00:00.