Hi Product Hunt! I’m Deonte Watts, the creator of Ethos Aegis. I started building Ethos Aegis around a question: What if artificial intelligence could defend itself using principles inspired by the human immune system? Biological immune systems don’t rely on a single defense. Different cells identify threats, communicate signals, preserve memory, and respond through specialized mechanisms. Ethos Aegis translates that idea into a developer-oriented AI security architecture. The project includes specialized defense components for input inspection, suspicious instruction detection, Unicode normalization, signature memory, and threat adjudication. It also brings five established security tools into a unified evidence-review workflow: TruffleHog: Credential-exposure evidence. Sherlock: Username-exposure observations. Ghidra: Static binary-analysis summaries. mitmproxy: Application-network observations. ESP32 Marauder: Operator-controlled wireless laboratory evidence. These integrations use explicit scope controls. External tool execution is not automatically enabled, and imported observations are not treated as independently verified vulnerabilities. The project also includes an interactive 4D threat-visualization board and recorded walkthroughs demonstrating the AI immune pipeline, security-evidence handling, and VeriFlow reasoning. The visual board is a simulation, while recorded walkthroughs use labeled synthetic inputs and actual local program output. I’m sharing Ethos Aegis with developers, security researchers, and AI builders who want to explore more transparent, evidence-driven security architectures. I’d especially appreciate feedback on the developer experience, security boundaries, and which integrations would be most useful next. Thanks for exploring Ethos Aegis! — Deonte Watts

Ethos Aegis

AI immune defenses for safer agents and security reviews

AI immune defenses for safer agents and security reviews