Content Guidelines

In short: share prompts that help others – and nothing that harms. When in doubt, the spirit of these rules applies, not their letter.

This is a convenience translation. The German version is legally binding.

Not allowed

• Harmful prompts: instructions or prompts for creating malware, for fraud, for generating illegal content, or for harming people.
• Jailbreaks and circumvention: prompts whose purpose is to bypass AI models' safety measures. We follow Anthropic's Usage Policy as our reference.
• Rights violations: content infringing copyright, trademarks, or personality rights – only publish what you hold the rights to.
• Personal data and secrets: no third-party data, no API keys, passwords, or internal company documents in prompts or examples.
• Spam: mass-produced, empty, or purely promotional repositories.

Expected

• Respect licenses: forks keep their origin visible; the original's chosen license continues to apply.
• Honest documentation: a prompt's known weaknesses belong in the repo – here that's a feature, not a flaw.
• Respectful conduct in discussions and reviews.

Reports and enforcement

You can report violations informally to benedikt@promthub.org (a link to the repo is enough). Reported content is reviewed and removed if in violation; repeated or severe violations lead to account suspension.

Note

PromptHub is an independent community project and is not affiliated with Anthropic.