Contribute

Anything you use regularly and would miss if it disappeared is worth publishing here.

What makes a good submission

Say where it was tested. "Works on Claude and GPT, unchanged" is more useful than a claim that it works everywhere.

Explain the clause that does the work. Most prompts have one line that is the reason they work. Point at it in the usage notes.

Record what breaks. Where a prompt fails - a smaller model, a longer input, a different language - is the part nobody else can write for you.

Keep placeholders obvious. Use {{double_braces}} and describe each one.

What gets rejected

Prompts that are only a role line. Configuration files copied from documentation without changes. Anything that cannot be tried by the reader because the important part is missing.

Editing

You can edit anything you published. Edits to a published resource create a new version and the previous one stays in the history, so links people saved keep working.