How m524security/sentrix works
SENTRIX is an autonomous backup-recovery readiness system: it ingests fleet state and telemetry, reasons over risk and history, predicts near-term breaches, scores recoverability from decaying restore evidence, and takes or recommends action. The agent publishes cycles to an API that serves operational views to a React dashboard. The graph preserves the documented ingestion, memory, recovery-evidence, forecasting, safety, and dashboard capabilities. Some subsystem links are documented rather than shown in sampled implementation excerpts; those are marked dashed.
Main entry points: agent/main.py Agent cycle · api/main.py Control API · reasoning_core.py Risk reasoning · tool_agent.py Tool investigation
Components 20
Fleet ingestion
- Fleet sources
agent/ingestion/fleet_source.py - LogHub engine
agent/ingestion/loghub_engine.py - Telemetry gateway
agent/ingestion/realtime_gateway.py - Telemetry adapters
agent/ingestion/adapters
Agent reasoning
- Agent cycle
agent/main.py - Risk reasoning
agent/reasoning/reasoning_core.py - LLM providers
agent/reasoning/llm_providers.py - Decision memory
agent/memory/decision_memory.py - Incident recall
agent/memory/incident_rag.py - RPO forecasting
agent/reasoning/predictive_engine.py - Tool investigation
agent/agency/tool_agent.py
Recovery evidence
- Recovery confidence
agent/recovery/confidence.py - Evidence ledger
agent/recovery/evidence_ledger.py - Restore scheduler
agent/recovery/evidence_scheduler.py
API and actions
- Action executor
agent/actions/executor.py - Control API
api/main.py - Operational persistence
agent/persistence.py
Dashboard
- Dashboard views
dashboard/src/App.js - Dashboard API client
dashboard/src/lib.js
Other components
- Engineer
Connections 23
- Agent cycle → Fleet sources: loads fleet
agent/main.py - Fleet sources → LogHub engine: uses simulation data
agent/ingestion/fleet_source.py - Engineer → Telemetry gateway: sends telemetry
README.md - Telemetry gateway → Telemetry adapters: parses telemetry
README.md - Agent cycle → Risk reasoning: reasons over assets
agent/main.py - Risk reasoning → LLM providers: requests model output
agent/reasoning/reasoning_core.py - Risk reasoning → Decision memory: recalls history
agent/reasoning/reasoning_core.py - Tool investigation → Decision memory: uses decision history
agent/agency/tool_agent.py - Tool investigation → Incident recall: recalls incidents
agent/agency/tool_agent.py - Agent cycle → Tool investigation: refines assessments
agent/main.py - Agent cycle → RPO forecasting: forecasts breaches
agent/main.py - RPO forecasting → LogHub engine: reconstructs history
agent/reasoning/predictive_engine.py - Agent cycle → Recovery confidence: scores readiness
agent/main.py - Recovery confidence → Evidence ledger: reads restore evidence
- Agent cycle → Restore scheduler: plans restore tests
agent/main.py - Restore scheduler → Recovery confidence: uses confidence model
agent/recovery/evidence_scheduler.py - Agent cycle → Action executor: dispatches actions
agent/main.py - Agent cycle → Control API: publishes cycles
agent/main.py - Control API → Operational persistence: stores cycle data
api/main.py - Control API → Fleet sources: serves fleet state
api/main.py - Dashboard views → Dashboard API client: uses API helpers
dashboard/src/App.js - Dashboard API client → Control API: requests API data
- Engineer → Dashboard views: views fleet readiness
README.md