Sondear Hardware Real
Ejecuta sondeo ICMP/TCP en tiempo real en la red del campus. Identifica switch de administración y APs caídos.
No ejecutado aún
DISPOSITIVOS SUPERVISADOS
106 total
14 Bellini · 92 Casa Lab
ESTADO RED BELLINI I-II
7 Online 7 Offline
Switch Admin & APs caídos
INFERENCIA LOCAL SRE
Qwen 30B AMD 1.5
Gasto: $0.00 (Local Invariant)
INCIDENTES ACTIVOS
P1 1 Crítico P2 1 Lab
Gestión en tiempo real
OPERACIONES VERIFICADAS / PRUEBA DE INFRAESTRUCTURA REAL
Operaciones verificables sobre dispositivos reales.
Conexión directa al nodo InnerOS AMD 1.5 mediante túnel cifrado inverso. Consulta hardware real, ejecuta inferencia local con Qwen 30B ($0.00 de gasto) y registra evidencias criptográficas SHA-256 inmutables. Cero estados fingidos ni simulaciones encubiertas.
Verificando servidor…
Diagnosticar con Qwen 30B
Inferencia local vLLM en AMD 1.5 ($0.00). Deducción determinista de causa raíz y protocolo de contención.
Esperando paso 1
Crear Ticket Operativo
Registra ticket formal P1 en MongoDB (ops_tickets) con ID inmutable, SHA-256 y clasificación de impacto.
Esperando diagnóstico
Notificar y Registrar ACK
Despacha alerta interna al lead de guardia (rafael) y registra acuse técnico humano (cero auto-ACK).
Esperando ticket
Recheck de Hardware
Vuelve a sondear equipos. Confirma persistencia física y rehúsa cierre fraudulento (Zero-Fabrication).
Esperando ticket
Seleccione el sitio y pulse 'Ejecutar sondeo real' para iniciar el triaje de hardware.
Telemetría de Hardware Real
Bellini Network Guardian · AMD 1.5
VERIFIABLE AUTONOMY
Agents that do the work.
Evidence that proves it.
From a real business incident to a documented, traceable outcome. Policy-driven workflows, isolated code execution, and tamper-evident logs.
WORKFLOW ENGINE
Checking…
Server-side orchestration
EXECUTION SANDBOX
Checking
Disposable, no-network containers
INFERENCE ENGINE
Checking
Qwen local / rules fallback
VERIFIED RUNS
0
Persistent evidence ledger
01 / ACTUAL WORK
Execute a workflow
AVAILABLE NOW
Incident → Action Plan
Turn a customer's infrastructure issue into a classified priority, an actionable service plan, and a SHA-256 evidence trail.
Public-demo submissions are visible to other visitors. Use fictional examples only.
DOCKER REQUIRED
Issue → Verified Repair
Reproduce a bug, generate a candidate patch, rerun assertions inside an ephemeral container, and record the outcome.
Untrusted code is never executed in the web app process.
02 / SECOND TECHNICAL FRONT
Containment Lab. A proof you can challenge.
Run one of three fixed safety probes in a disposable Docker container. The backend records its real outcome and independently checks container removal. No worker attached? The test is marked BLOCKED, not silently simulated.
This is a technical safety demonstration, not a guarantee against container escape. Blast Radius Zero still requires Vultr Serverless Inference and execution on Vultr.
Checking isolated worker…
No containment test run yet.
LIVE RESULT
Execution
EVIDENCE ARTIFACT
—
LAST EVIDENCE HASH
—INTEGRITYVerification pending
03 / AUDITABILITY
Execution ledger
WORKFLOWSTATUSEVIDENCE
No executions yet. Run a workflow to create a trace.
04 / CASE LIFECYCLE
Service ticket board
No service tickets yet. Run incident triage to open one.
05 / THE CONTROL PLANE
Build for proof, not promises.
01 · PLAN
Classify business impact and define a bounded task.
02 · EXECUTE
Send risky code into short-lived isolated containers.
03 · VERIFY
Test results and capture actual execution status.
04 · ATTEST
Append a SHA-256-linked evidence trail.