Prompts
Prompt testing needs more than a clever paragraph copied from a chat window. This category publishes deterministic creator briefs, vendor-neutral multi-step plans, OpenAI-compatible message and tool-call examples, Anthropic-compatible message examples, and evaluation contracts with explicit expected results. Fifty cross-category workflow kits connect each brief and plan to real input media, a reference output, and a controlled edge case through one stable group ID. Provider labels describe syntax compatibility rather than endorsing a service, and every text fixture can be inspected before download.
Filter prompts on Browse · 300 files · 4 subcategories
Frequently asked questions
Are these prompts tied to one AI provider?+
Most are vendor-neutral plans with explicit inputs and expected outcomes. Separate OpenAI-compatible and Anthropic-compatible examples exercise their message shapes without requiring an account or live API call.
How do the workflow kits connect prompt files to media?+
Each kit uses one stable group ID and six ordered roles: creator brief, prompt plan, evaluation contract, input media, expected output, and a controlled edge case.
Do the examples include secret keys or real user data?+
No. Credentials are placeholders, names and scenarios are invented, and every artifact is generated locally from a fixed recipe.