A next-generation honeypot system that uses a five-layer cognitive architecture to create adaptive, intelligent deception environments. Mirage transforms traditional honeypots from static decoys into dynamic, learning systems that adapt to attacker behavior in real-time.
Primary Metric: Mean Time To Discovery (MTTD)
Current Baseline: 2-5 minutes (static honeypot)
Target Goal: 45-60+ minutes (9-12x improvement)
For a deep dive into the system architecture and implementation details, please refer to the 🏗️ Technical Foundations.
- 📚 Usage Guide - Setup and operation
- 📊 Progress Tracking - Development status
- 🧠 AI Engine Plan - Cognitive roadmap
- 🔧 API Reference - API documentation
- Layer 0 - Reflex Layer: Sub-millisecond deterministic threat detection in Rust
- Cognitive Pipeline: Cascading short-circuit logic to route traffic efficiently
- Layer 1 - Intuition Layer: Real-time command prediction using Hidden Markov Models
- Layer 2 - Reasoning Layer: Attacker behavioral classification with Machine Learning
- Layer 3 - Strategy Layer: Long-term engagement optimization via Reinforcement Learning
- Layer 4 - Persona Layer: Context-aware conversational responses using LLMs
- Predictive Modeling: Anticipates attacker actions before they happen
- Behavioral Learning: Builds comprehensive attacker profiles over time
- Strategic Optimization: Learns optimal deception strategies through self-play
- Dynamic Personas: Maintains consistent character across extended interactions
Foundation Complete: 87% ✅
Mirage Architecture: 3% (Layer 0 in progress)
| Layer | Component | Status | Target Timeline |
|---|---|---|---|
| Foundation | Apate Core (SSH/HTTP/DB) | ✅ Complete | - |
| Layer 0 | Reflex Layer (Rust) | 🔄 In Progress | Q4 2025 |
| Layer 1 | Intuition Layer (HMM) | ⏳ Planned | Q1 2026 |
| Layer 2 | Reasoning Layer (ML) | ⏳ Planned | Q2 2026 |
| Layer 3 | Strategy Layer (RL) | ⏳ Planned | Q3 2026 |
| Layer 4 | Persona Layer (LLM) | ⏳ Planned | Q4 2026 |
| Phase | Layers Active | Target MTTD | Improvement | Timeline |
|---|---|---|---|---|
| Baseline | Static Foundation | 2-5 min | 1x | Current |
| Phase 1 | Layer 0+1 | 15-20 min | 3-4x | Q1 2026 |
| Phase 2 | Layer 0+1+2 | 25-35 min | 5-7x | Q2 2026 |
| Phase 3 | Layer 0+1+2+3 | 35-50 min | 7-10x | Q3 2026 |
| Phase 4 | All Five Layers | 45-60+ min | 9-12x | Q4 2026 |
Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests.
This project is licensed under the MIT License - see the LICENSE file for details.
Note
Work in Progress: The AI components (Layers 1-4) described in this document are currently in the planning and development phase. The current codebase reflects the robust static foundation and the initial implementation of Layer 0 (Reflex Layer). Please check back for updates as we implement the cognitive architecture.