discovered 11 Aug 2026
ai-red-teaming
→ View on GitHubRed-Team AI is a white-box red teaming tool designed specifically for agentic AI applications, capable of reading source code to identify vulnerabilities that are unique to a particular technology stack. Its primary use case involves generating tailored attacks based on an application's specific implementation, rather than relying on generic adversarial prompts. Notable features include a modern React dashboard for scan management and compliance tracking, as well as integrations with popular agent frameworks, facilitating extensive security assessments and risk assessments for AI systems.