FSI Vitalis CyberCore
Ferrell Synthetic Intelligence (FSI)
Built by Neuro_Nomad
What This Is
A demonstration fork of Vitalis Core trained toward cybersecurity threat detection and classification.
This shows developers exactly what Vitalis Core becomes when directed toward a specific domain. Clone the base model. Train it your way. Own it completely.
What It Does
- Classifies input as HIGH, MEDIUM or LOW severity threats
- Identifies defensive actions and training signals
- Assigns confidence scores to every classification
- Logs everything to an immutable JSON-Lines audit trail
- Config-driven — add keywords without touching code
- Zero network calls. Fully air-gapped capable.
Benchmark Results
malware detected → THREAT_DETECTED [HIGH] — Confidence: 0.90 unauthorized access → THREAT_DETECTED [MEDIUM] — Confidence: 0.85 port scan detected → THREAT_DETECTED [LOW] — Confidence: 0.70 monitor the network → DEFENSIVE_ACTION — Confidence: 0.60 harden the firewall → DEFENSIVE_ACTION — Confidence: 0.60
All tests passing.
Run It
git clone https://huggingface.co/FerrellSyntheticIntelligence/FSI-Vitalis-CyberCore cd FSI-Vitalis-CyberCore python3 fsi_main.py
Run Tests
python3 tests/test_brain.py
Extend It
Edit core/config.yaml to add your own threat keywords. No code changes required.
Base Model
Built on Vitalis Core. huggingface.co/FerrellSyntheticIntelligence/Vitalis_Core
Who Built This
Self-taught. No degree. No team. No funding. FSI — Ferrell Synthetic Intelligence Handle: Neuro_Nomad
License
GPL-3.0
- Downloads last month
- 23