Inference-Based Intrusion Detection for AI Agents
"Traditional IDS matches signatures. HoPE understands intent."
Goes beyond pattern matching. Uses semantic analysis to understand what an attack is trying to do, catching novel and obfuscated threats.
Heuristic β Semantic β Context β Decision β Learning. Fast pattern matching plus deep analysis when needed.
Pure Node.js. No cloud services required. Works offline with heuristic mode, or connect any OpenAI-compatible LLM for semantic analysis.
HoPE-voiced alerts that are informative and memorable. "Nope. 'Ignore previous instructions' doesn't work on me. I know who I am."
Tracks sender history, source trust levels, rate patterns. Adjusts risk based on who's talking and where the message came from.
CLI tool, middleware, Docker container. Integrate in minutes with Express, Hono, or OpenClaw.
Real detection results from hopeIDS