When Anthropic and OpenAI started publishing the system prompts behind their assistants, something quietly important happened: the instructions that shape how a model behaves stopped being a secret. You could read them, compare them, and learn from them.
OpenPrompts exists to gather that openness in one place, across providers, across types. System prompts, skills, and ready-to-use prompts, each with its source and license kept intact.
Read the instructions
A hidden prompt can't be reviewed. It can't be criticized, improved, or trusted. Reading prompts in the open is how we hold systems, and the people who build them, accountable.
That's the whole idea. Browse the catalog, copy what's useful, and if you find a prompt the community should see, add it.