# asgeirtj/system_prompts_leaks

> A collection of extracted system prompts from Anthropic, OpenAI, Google, xAI and others — the hidden instructions chatbots receive before a user's first message.

- Magnitude: 2.7 out of 10 — Cooling
- Stars: 68,372 total · +4 stars on 2026-09-26 (UTC), ≈ 31 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: JavaScript · License: CC0-1.0 · Created: 2025-05-03 · Last push: 2026-09-22
- GitHub: https://github.com/asgeirtj/system_prompts_leaks · Page: https://gitnova.dev/en/r/asgeirtj/system_prompts_leaks

## Useful for

- Study Claude Code system prompts to understand how the agent plans file operations
- Compare ChatGPT and Gemini instructions to analyze differences in model behavior
- Use the prompts as a basis for writing your own agent instructions

## Why it’s here

- 4 stars so far today, about 31 expected by the end of the day.
- Interest is fading: the two-day pace is 24% of the previous week and a half.
- About 14 forks a day — people are taking the code.
- Recent forks include notable developers: @felipealvessi (227 followers).

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 11,111
- Issues and pull requests: 189
- Watchers: 768
- Average over the last week: 103 per day
- Usual pace: 216 per day

## Stars per day, last 30 days (oldest → newest, today is partial)

2026-09-30 … 2026-10-29: None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None

## Magnitude by day, last 14 days

09-13 5.7, 09-14 6.7, 09-15 6.1, 09-16 4.4, 09-17 3.1, 09-18 2.9, 09-19 3.6, 09-20 3.7, 09-21 3.2, 09-22 2.9, 09-23 3.0, 09-24 3.1, 09-25 3.3, 09-26 2.9

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-09-26 10:58 UTC, updated every 30 minutes.