Self-Reflective Autonomous Multi-Agent Security Swarm with Deterministic Cryptographic Perimeter Enforcement
This paper presents the system architecture and empirical cyber-range validation of BAGLEY-AI, an autonomous on-premise multi-agent security engine implemented in compiled Rust and C++. Addressing the dual risks of source-code exfiltration in cloud scanners and behavioral drift in autonomous agents, we introduce a dual-layer deterministic perimeter enforcement model (Layer 1 Kernel Interceptor and Layer 2 Isolated Behavioral Sentinels) backed by an immutable SHA-256 hash-chained audit ledger.
Authors
- BAGLEY-AI Security Research
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-19
- DOI
- https://doi.org/10.5281/zenodo.24371852
- Primary Topic
- Security and Verification in Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00