stylusnexus/agent-armor
Open-source agent security framework. Detects and defends against AI Agent Traps — content injection, embedded jailbreaks, RAG poisoning, data exfiltration, and more. Based on the DeepMind Agent Traps taxonomy.
GitHub repository with 5 stars and 0 forks.
Language: TypeScript
Topics: agent-safety, agent-security, ai-agents, ai-governance, ai-security, audit-logging, compliance, jailbreak-detection, llm-security, onnx