How spod-codes/cyberoracle works
CyberOracle is a network-telemetry threat-analysis application: analysts upload CSV data or monitor enrolled agents, then review risk, forecasts, explanations, and operational summaries. The main offline workflow runs from the evaluation API through the world model to a stored evaluation that the frontend presents in dashboard and briefing views. Live supervision is a separate ingestion path. Source excerpts show backend routing and model responsibilities, but frontend API wiring and several live-processing details were not sampled; those links are omitted or marked unverified.
Main entry points: App.tsx Page router · api.ts API client · evaluations.py Evaluation routes · world_model.py World model
Components 24
Analyst interface
- Page router
frontend/src/App.tsx - Upload home
frontend/src/pages/Home.tsx - Risk dashboard
frontend/src/pages/Dashboard.tsx - Incident briefing
frontend/src/pages/Briefing.tsx - Live monitor
frontend/src/pages/LiveMonitor.tsx - Traffic simulator
frontend/src/pages/Simulator.tsx - API client
frontend/src/lib/api.ts
Offline analysis
- API server
backend/server.py - Evaluation routes
backend/routers/evaluations.py - World model
backend/lib/world_model.py - Feature extraction
backend/lib/features.py - Sequence forecasting
backend/lib/sequence_models - Model explanations
backend/lib/explainability.py - MITRE classification
backend/lib/mitre_classifier.py - Kernel evaluation CLI
backend/kernel_cli.py
Live supervision
- Live routes
backend/routers/live.py - Flow aggregation
backend/lib/flow_aggregator.py - Agent supervision
backend/routers/supervision.py - Endpoint agent
backend/lib/agent_script.py - Risk calibration
backend/lib/calibration.py
Persistence and contracts
- Evaluation records
backend/models/evaluation.py - Mongo database
backend/lib/db.py - Supervision records
backend/models/supervision.py
Other components
- Analyst
Connections 30
- Analyst → Page router: uses
- Page router → Upload home: routes to
frontend/src/App.tsx - Page router → Risk dashboard: routes to
frontend/src/App.tsx - Page router → Incident briefing: routes to
frontend/src/App.tsx - Page router → Live monitor: routes to
frontend/src/App.tsx - Page router → Traffic simulator: routes to
frontend/src/App.tsx - Upload home → API client: submits data
- Risk dashboard → API client: fetches results
- Incident briefing → API client: updates review
- Live monitor → API client: reads live data
- Traffic simulator → API client: sends traffic
- API client → API server: calls API
- API server → Evaluation routes: registers
backend/server.py - API server → Live routes: registers
backend/server.py - API server → Agent supervision: registers
backend/server.py - Evaluation routes → World model: runs analysis
backend/routers/evaluations.py - Evaluation routes → Mongo database: reads and writes
backend/routers/evaluations.py - Evaluation routes → Evaluation records: returns records
backend/routers/evaluations.py - World model → Feature extraction: extracts features
backend/lib/world_model.py - World model → Sequence forecasting: forecasts trajectory
backend/lib/world_model.py - World model → Model explanations: computes explanations
backend/lib/world_model.py - World model → MITRE classification: classifies tactics
backend/lib/world_model.py - Evaluation routes → Kernel evaluation CLI: launches tests
backend/routers/evaluations.py - Live routes → Flow aggregation: aggregates flows
backend/routers/live.py - Live routes → Model explanations: requests explanations
backend/routers/live.py - Live routes → Sequence forecasting: uses forecasters
backend/routers/live.py - Agent supervision → Mongo database: stores agent data
backend/routers/supervision.py - Agent supervision → Endpoint agent: serves agent
backend/routers/supervision.py - Agent supervision → Risk calibration: proposes adjustments
backend/routers/supervision.py - Agent supervision → Supervision records: validates records
backend/routers/supervision.py