@XAMTO_AI: Regarding the leaked system prompt for Claude Fable 5, someone created a token-optimized version and converted it to clean Markdown format! The code repository claims to have refactored it into a universal format that can be directly applied to frontier models such as Gemini 3.1 Pro and ChatGPT 5.6. htt…
Summary
Someone token-optimized the leaked Claude Fable 5 system prompt, converted it to clean Markdown format, refactored it into a universal format, and it is applicable to various frontier AI models.
View Cached Full Text
Cached at: 08/21/26, 05:05 AM
Regarding the leaked Claude Fable 5 system prompt, a token-optimized version has been created and converted into clean Markdown format! The repository claims it has been refactored into a universal format that can be directly applied to cutting-edge models like Gemini 3.1 Pro, ChatGPT 5.6, and others. https://github.com/KinetiNode/claude-fable-5-system-prompt-clean…
KinetiNode/claude-fable-5-system-prompt-clean
Source: https://github.com/KinetiNode/claude-fable-5-system-prompt-clean
Universal Fable 5 System Prompt
License: MIT (https://opensource.org/licenses/MIT) Compatibility
A vendor-neutral adaptation of Anthropic’s publicly available Claude Fable 5 / Mythos 5 system prompt.
The original prompt contains a large amount of infrastructure that exists specifically for Claude’s production environment, including XML conventions, tool interfaces, product-specific behavior, and internal implementation guidance. While appropriate for Claude, much of that content provides little or no benefit when reused with other frontier LLMs.
This repository filters out those implementation details while retaining the behavioral guidance that generalizes well across modern models.
Goals
- Reduce prompt size for better context efficiency.
- Remove Claude-specific implementation details.
- Preserve high-impact behavioral guidance.
- Maintain compatibility across major frontier LLMs.
Versions
| File | Purpose |
|---|---|
universal/core.md | Core behavioral principles with minimal token usage. |
universal/balanced.md | Recommended for most use cases. Balances prompt size with behavioral coverage. |
universal/complete.md | Retains nearly all vendor-neutral behavioral guidance while remaining model-agnostic. |
What’s Removed
This project intentionally excludes Claude-specific implementation such as:
- XML wrappers and formatting conventions
- MCP and tool-specific infrastructure
- Product-specific instructions
- Internal UI behavior
- Storage and artifact management
- Environment-specific implementation details
What’s Preserved
The filtered prompts retain behavioral guidance including:
- Understanding user intent before responding
- Honest handling of uncertainty
- Clear communication and natural writing
- High-quality code generation
- Practical recommendations with trade-off analysis
- Verification before finalizing responses
- Objective error correction
- Avoiding unnecessary meta-commentary
Compatibility
The prompts are designed to work well across modern frontier models, including:
- Claude
- ChatGPT
- Gemini
- DeepSeek
- Qwen
- Grok
- GLM
- Other instruction-following LLMs
Usage
API Platforms
Paste the contents of the desired prompt into your model’s system prompt or system instructions field.
Consumer Interfaces
Where supported, the prompts can also be used as custom instructions, custom GPT system prompts, Gems, or equivalent features.
Philosophy
This project is not intended to replace Anthropic’s original system prompt or claim to improve it.
Instead, it asks a different question:
Which behavioral principles from Claude’s system prompt remain valuable after removing implementation details that are specific to Claude itself?
The result is a set of prompts optimized for portability across platforms rather than platform-specific functionality.
Contributing
Contributions are welcome.
If you discover vendor-specific behavior that should be generalized, redundant instructions that can be simplified, or opportunities to improve compatibility across modern LLMs, feel free to open an issue or submit a pull request.
Similar Articles
I distilled the leaked Claude Fable 5 system prompt into a clean, universal 500-token Markdown engine for ChatGPT and Gemini. No bloat.
A developer distilled the leaked Claude Fable 5 system prompt into a compact 500-token universal Markdown engine for ChatGPT and Gemini, removing Anthropic-specific bloat and making it usable on any frontier model.
@GitTrend0x: Claude Code Codebase Smart Brain — 27× Token Savings Killer Open-Source Tool https://github.com/repowise-dev/repowise… This is Repowise, a codebase intelligence platform built for AI-assisted engineering teams! It turns...
Repowise is an open-source tool that indexes codebases into four intelligence layers (dependency graph, git history, auto-documentation, architectural decisions) and exposes them via seven MCP tools to AI coding agents like Claude Code, achieving up to 27× token savings while maintaining answer quality.
@axichuhai: Found an open-source tool that makes Claude even better — just a 65-line Markdown file, already reached 130k+ stars on GitHub. Inspired by Karpathy's observation that models like to guess when uncertain, write overly complex code, and quietly change things they shouldn't touch...
An open-source 65-line Markdown file that summarizes 4 programming rules for Claude to improve code quality and save tokens. It has garnered 130k+ stars on GitHub, inspired by Karpathy.
@sunmer575399: Recommending an open-source project I've been using recently, caveman, 77.3k stars. It can cut Claude Code conversation tokens by 65%, by making the AI reply in caveman grammar, saving money. I ran a week's worth of work, and the bill was literally halved. No need to change your workflow, just install it as Cla…
Recommending an open-source project, caveman, which reduces output tokens by making AI reply in caveman grammar. It can save 65% of prose tokens and 8.5% of agentic coding task tokens, and is compatible with 30+ AI agent tools.
@GYLQ520: Attention AI Agent developers! Token costs burning a hole in your pocket? There's an open-source tool called curl.md that converts web pages to Markdown format for AI, slashing token consumption significantly. Choose from CLI, browser extension, or API usage. Cursor, Claud…
curl.md is an open-source tool that converts web pages to optimized Markdown format for AI agents, significantly reducing token consumption and cost. It offers CLI, browser extension, and API usage, with integrations for Cursor, Claude, and other agents.