Master's Thesis¶
Historical thesis snapshot. This page describes Fernando Boiero's Fernando-only thesis snapshot for MIESC v4.0.0. It intentionally preserves the thesis-era 7-layer / 25-tool framing and thesis metrics. Current project and paper claims are maintained separately in the README, paper reproducibility notes, and claims matrices.
Thesis Information¶
| Field | Value |
|---|---|
| Title | Integrated Security Assessment Framework for Smart Contracts: A Defense-in-Depth Approach to Cyberdefense |
| Author | Fernando Boiero |
| Advisor | M.Sc. Eduardo Casanovas |
| Institution | Universidad de la Defensa Nacional (UNDEF) - IUA Cordoba |
| Program | Master's Degree in Cyberdefense |
| Expected Defense | Q4 2025 |
Abstract¶
This thesis presents MIESC (Multi-layer Intelligent Evaluation for Smart Contracts), a production-grade security framework that implements a 7-layer Defense-in-Depth architecture for comprehensive smart contract vulnerability detection.
The framework integrates 25 specialized security tools with AI-powered correlation using sovereign LLMs (Ollama) and ML-based detection (DA-GNN Graph Neural Networks), achieving 94.5% precision, 92.8% recall, and an F1-score of 0.93.
Key innovations include: - Triple normalization system (SWC/CWE/OWASP) with 97.1% accuracy - MCP Protocol integration for AI assistant interoperability - Sovereign AI backend ensuring data never leaves the user's machine - Legacy tool rescue for deprecated but valuable security tools
Chapters¶
Part I: Foundations¶
| Chapter | Title | Description |
|---|---|---|
| 1 | Introduction | Problem statement, objectives, and thesis structure |
| 2 | Theoretical Framework | Blockchain, smart contracts, and security fundamentals |
| 3 | State of the Art | Existing tools, frameworks, and research |
Part II: Implementation¶
| Chapter | Title | Description |
|---|---|---|
| 4 | Development | MIESC architecture, agents, and implementation details |
| 5 | Experimental Results | Benchmarks, metrics, and comparative analysis |
Part III: Justification¶
| Chapter | Title | Description |
|---|---|---|
| 6 | AI and Sovereign LLM Justification | Data sovereignty, Ollama integration, DPGA compliance |
| 7 | MCP Protocol Justification | Model Context Protocol, tool handlers, interoperability |
Part IV: Conclusions¶
| Chapter | Title | Description |
|---|---|---|
| 8 | Future Work | Research directions, planned enhancements |
Key Metrics¶
Framework Performance (v4.0.0)¶
| Metric | Value |
|---|---|
| Precision | 94.5% |
| Recall | 92.8% |
| F1-Score | 0.93 |
| False Positive Rate | 5.5% |
| Detection Coverage | 96% |
| Integrated Tools | 25 |
| Defense Layers | 7 |
| Compliance Index | 91.4% |
ML Detection (DA-GNN)¶
| Metric | Value |
|---|---|
| Accuracy | 95.7% |
| False Positive Rate | 4.3% |
| Graph Representation | CFG + DFG |
Research Contributions¶
-
7-Layer Defense-in-Depth Architecture - Novel multi-layer approach combining static, dynamic, symbolic, formal, AI, ML, and audit layers
-
25 Tool Integration - Unified ToolAdapter protocol for seamless tool interoperability
-
Triple Normalization System - SWC/CWE/OWASP mapping with 97.1% accuracy
-
Sovereign AI Backend - Ollama integration ensuring data sovereignty and $0 operational cost
-
MCP Server - Model Context Protocol for AI assistant integration
-
Legacy Tool Rescue - Manticore Python 3.11 compatibility, Oyente Docker containerization
Citation¶
@mastersthesis{boiero2025miesc,
author = {Boiero, Fernando},
title = {MIESC: Multi-layer Intelligent Evaluation for Smart Contracts},
school = {Universidad de la Defensa Nacional (UNDEF)},
year = {2025},
type = {Master's Thesis},
address = {Cordoba, Argentina},
note = {Master's Degree in Cyberdefense}
}