TECHNOLOGY & RESEARCH
The SenseL roadmap combines agent action traces, edge sensing, versioned evidence, constrained AI reasoning, reproducible evaluation, and controlled validation. Every capability carries an explicit maturity label and a clear current boundary.
TECHNOLOGY & RESEARCH
Models change, data goes missing, and sites go offline. Our technical strategy is to preserve provenance, versions, gaps, and authority first; accelerate investigation inside those boundaries; then use safe scenarios and retesting to prove whether an improvement worked.
EDGE → EVIDENCE → VALIDATION
Every layer retains a clear owner, version, and failure state. AI cannot manufacture detector values or turn missing evidence into fact.
01
Agent, Tool, MCP, edge, EDR, OT, and CTI signals
02
Provenance, versions, gaps, and evidence IDs
03
Constrained AI summaries and investigation hypotheses
04
Frozen data, manifests, and reproducible evaluation
05
Isolated scenarios, human approval, and guardrails
06
Remediation, retest, and versioned improvement records
BUILDING IN PUBLIC
These programs advance the SenseL roadmap, but not every item is a generally available product feature. Availability varies by maturity, deployment path, and technical review.
EDGE + OT INTELLIGENCE
Engineering Preview · Public RepoCombines active EdgeX telemetry with passive SPAN/TAP sensing at the edge, sending only feature summaries, security events, and evidence references northbound.
Public MVP: OT rules, IEC 61850, MQTT, and Suricata/Snort bridges. SenseL OT Dashboard integration remains on the roadmap.
View public repositoryEVIDENCE CONTRACT
Research Preview · Pre-submissionResearch into packaging EDR/NDR investigations as versioned, replayable, auditable evidence contracts that keep missing evidence missing instead of letting AI fill the gap.
Public research artifact. It does not imply paper acceptance or full integration of every contract into the production SenseL platform.
View research artifactREPRODUCIBLE EVALUATION
Evaluation Artifact · R0/R1Uses frozen manifests, fixtures, checksums, and separated Safe and Attack-enabled conditions to test whether AI-assisted security validation can be independently recomputed.
Public R0 artifact verification and R1 fixture recomputation are available. R2 live execution remains behind an authorized lab gate.
View reproducible evaluationSAFE OT VALIDATION
Research Preview · Pre-submissionResearch on turning IEC 61850 threats and CVEs into controlled validation hypotheses while preserving MMS/GOOSE baselines and fail-closed safety boundaries.
Currently pre-submission research. Event counts do not establish exploit success or physical relay operation.
View research packageHEALTHCARE EDGE AI
Research Preview · Private LabLocal-first passive IoMT/OT security research: detectors own the facts, a constrained small language model formats evidence, and deterministic fallback takes over when generation fails.
Advisory only. It has not completed clinical-use, production hospital, or medical-device regulatory validation.
VALIDATION LAB
Lab Capability · ARM-ready BaselineAn isolated, resettable defender environment combining Wazuh, CALDERA, safe simulation, and local AI-assisted evidence review for training and validation.
ARM-ready lab baseline. x86 OVA, 30 physical endpoint capacity, and disconnect recovery remain future validation gates.
CLAIM DISCIPLINE
Transparent maturity is not a disclaimer; it is part of technical trust. Before a capability becomes part of an offering, it should have reproducible evidence, explicit boundaries, and a defined promotion gate.
01
An executable engineering baseline or public MVP exists, while interfaces, integration, or field adaptation may still change.
02
Public data and tooling can verify a scoped evaluation claim; this is not the same as a product release or customer deployment.
03
Used to test a method or hypothesis. A submission, lab result, or prototype is not presented as a commercial capability.
04
Scenarios and evidence capture work in an isolated environment; field scale and production operations require separate acceptance.
FOUNDER-LED TECH REVIEW
The SenseL roadmap combines agent action traces, edge sensing, versioned evidence, constrained AI reasoning, reproducible evaluation, and controlled validation. Every capability carries an explicit maturity label and a clear current boundary.
