PromptRedTeam

What PromptRedTeam is

PromptRedTeam is a tool for scanning user input for attacks before it reaches a large language model (LLM). It detects direct injection, role manipulation, zero-width injection, delimiter injection, and encoded payloads. Its checks include pattern matching for override keywords, instruction conflicts, and prompt structure, as well as recognition of jailbreak patterns, role-play scenarios, and bypass attempts. It can detect zero-width characters, normalize Unicode, inspect Base64 and other encodings, and decode suspicious text before scanning again. Delimiter checks look for context-breaking markers, Markdown abuse, and structural manipulation. Applications can use its hosted API, which the homepage illustrates with a POST request to https://api.promptredteam.com/test, or self-host the project. The live demo accepts up to 1,000 characters and allows 10 requests per minute; self-hosting provides unlimited usage. Python, JavaScript, and cURL examples are available. The suggested workflow is to scan inputs before they reach an LLM, then block threats or log them for review. PromptRedTeam is open source under the MIT License, with web, API, and self-hosted options.