Kimi K2.5 System Prompt Leaked on GitHub (5k tokens)
Kimi K2.5's system prompt has been leaked on GitHub, revealing approximately 5,000 tokens of instructions that guide the AI model's behavior, responses, and
Someone managed to extract the complete system prompt from Moonshot’s Kimi K2.5 and dropped it on GitHub before most people even got to it.
What’s in there:
- Full system prompt (~5k tokens)
- All tool schemas and instructions
- Memory CRUD protocols
- Context engineering setup
- User profile assembly logic
- Basic guardrails
- External datasource configs (finance, arxiv, etc)
The repo is at https://github.com/dnnyngyen/kimi-k2.5-prompts-tools
They verified it across multiple accounts and got independent confirmation from several sources. Pretty rare to see someone grab this stuff so early - usually these prompts leak way later. Turns out Kimi has some interesting architecture around memory management and context assembly that isn’t documented anywhere else.
Good reference if you’re building similar systems or just curious how production LLMs structure their instructions behind the scenes.
Related Tips
"Take a Deep Breath" Boosts AI Accuracy on Hard Tasks
Research reveals that adding the phrase 'take a deep breath' to AI prompts significantly improves performance on complex reasoning tasks by encouraging more
LLMs Can Now Play Balatro Autonomously via API
An article discusses how large language models have gained the ability to autonomously play the poker-themed roguelike deck-building game Balatro through API
ACE-Step v1 Runs on 8GB VRAM with CPU Offload
ACE-Step v1 demonstrates efficient AI model execution on consumer hardware by running on systems with only 8GB VRAM through CPU offloading techniques that