Inspect the evidence
Open an incident, follow its source alert links, and export an incident report. Operators can link an existing response approval with an audit reason.
AKLupX is a pre-release Linux firewall appliance with network threat detection and evidence-linked AI incident review in a private operator console. Review alerts, track response decisions, and send telemetry securely to AWS with retry after disconnection.
// Capabilities
The current build brings filtering, event storage, and alert review together. MyNew Technologies has verified these features in controlled lab tests and completed a limited internal pilot.
Apply local filtering rules and review which connections the firewall allowed or blocked.
Review detections and sensor health in the private console.
Investigate detections, record approval decisions, and export alert reports as CSV.
Keep alerts and sensor health records within configured storage and retention limits.
Enroll devices with certificates, queue retained events locally, and retry uploads when connectivity returns.
AKLupX blocked controlled test traffic on an isolated path and restored forwarding after rollback.
Phase 9 · Lab verified
AKLupX now groups related network alerts and uses an AWS-hosted model to prepare advisory summaries for the private operator console. Each incident keeps its source alerts, confidence label, model identifier, and prompt fingerprint available for review.
Open an incident, follow its source alert links, and export an incident report. Operators can link an existing response approval with an audit reason.
Model summaries are hypotheses for human review. Approval decisions and firewall operations remain explicit operator actions.
A private model connection, certificate-based AWS access, bounded inputs, and daily call limits constrain the lab service. Analysis failures remain visible while telemetry continues.
Verified November 21, 2026: six incidents generated from eleven lab alerts, source references checked, and telemetry continued with the reasoning service stopped and restarted. Tests included synthetic evidence and a hostile instruction embedded in an alert. This is a limited internal lab result; broader client validation is ahead.
Threat context
ARTEX is an open-source, AI-powered penetration-testing framework that automates offensive security workflows. Microsoft has reported attackers using it to accelerate reconnaissance, password spraying, and abuse of trusted cloud applications.
Our goal is to give MyNew Technologies and its clients an inspectable path from local detection to an informed response: Linux filtering and passive sensing, resilient AWS telemetry, and evidence-led operator review.
The current lab combines network alerts, telemetry recovery, human approval records, and advisory AI incidents. The internal pilot report below documents the earlier passive traffic and recovery tests.
Our ARTEX-informed roadmap adds identity signals and cross-source investigation to address attacks that can move through legitimate cloud access.
Read Microsoft’s ARTEX research. Our lab tests establish the capabilities described here; ARTEX-specific identity defense requires the planned integrations and validation.
Phase 10 · Live sources verified
AKLupX now collects read-only Entra directory audit and user sign-in records and delivers selected evidence through authenticated AWS telemetry to the private operator console. Deterministic checks flag credential changes, application consent, privilege assignments, and repeated failed sign-ins for human review.
Each identity finding links its supporting source events and rule version. Replay records remain visibly labeled. An administrative change can be legitimate; a review finding alone does not establish compromise.
Ten queued replay events survived a scoped HTTPS interruption and arrived after connectivity returned. Automatic recovery took about sixteen seconds; duplicate delivery created no extra findings.
The console distinguishes live Microsoft source records from synthetic replay and shows collection health and freshness. Findings retain their supporting evidence and rule version; the collector has read-only audit access.
Verified November 21, 2026: real Entra directory audit and user sign-in records reached the private console through AWS. Separately, ten synthetic source events produced four review findings, and queued delivery recovered after an HTTPS interruption. These limited lab results do not demonstrate ARTEX detection, identity attack prevention, comprehensive workload-identity coverage, or Microsoft 365 mailbox monitoring.
Phase 11 · Isolated lab verified
Two inert file canaries now report observed reads through authenticated AWS telemetry into the private Deception and Incidents views. Each finding links its source evidence and detector version for operator review.
Four authorized, isolated content reads produced four alerts and three incidents. Repeated reads of the same canary share an incident within a watch session; a restarted watcher begins a new session.
The test reader had no external network route or access to real credentials. An ordinary-file read and metadata-only check produced no canary events. Decoy values are inert and stay out of telemetry.
A queued canary event survived a twenty-second publisher pause and arrived after restart. A controlled reboot preserved all four events and restored the watches, private relay, and network and identity telemetry.
Verified November 21, 2026 in our existing lab. High confidence refers to an observed file read; it does not identify the reader or establish compromise, cloud credential use, ARTEX detection, or prevention. These file canaries do not cover every access type. Broader cloud canaries, client validation, and Phase 12 correlation and notifications remain planned.
Next · Phase 12 and broader validation
Phase 9 provides advisory network analysis. Phase 10 adds verified live Entra evidence and private identity review. Phase 11 adds isolated file-canary evidence. Next comes cross-source correlation, notifications, and broader validation.
Measure detection quality against authorized tenant changes, extend workload-identity coverage, and test longer passive observation windows before client deployment.
Phase 11 verified two inert file canaries, four isolated reads, and delivery into private alerts and incidents. Broader cloud and OAuth canaries remain future work.
Phase 12 plans identity and network correlation with notifications. Operators review the evidence and decide how to respond.
Comprehensive workload-identity monitoring, impossible-travel detection, and client-scale validation remain future work. Demonstrations will use an agreed scope and clearly labeled evidence.
On November 17, 2026, MyNew Technologies passively observed one HTTPS path on its own server and separately tested AWS telemetry recovery after a simulated uploader outage.
These pre-release results cover one passive capture and one short outage. Production inline enforcement requires separate validation.
// About
Serious cybersecurity has stayed complex and expensive while the businesses that need it have gotten smaller and faster. AKLupX closes that gap.
AKLupX combines a Linux firewall, passive network detection, and evidence-linked AI incident review in a private operator console. Operators can review detections, record decisions, and export alerts.
The pre-release build has passed controlled tests for telemetry recovery, alert review, and isolated inline blocking and rollback. A November 2026 internal pilot verified passive observation on one owned server and AWS telemetry recovery after a simulated outage.
MyNew Technologies is a Chicago-based technology firm that builds and operates scalable systems across web, commerce, and infrastructure. Working in the Linux, Apache, MySQL, and PHP stack, the team designs custom, API-driven solutions engineered for resilience.
MyNew Technologies builds and tests AKLupX for small businesses and teams that want clearer network alerts and a record of how they responded.
// Pre-Release
AKLupX is in pre-release. Contact MyNew Technologies for a preview or to ask about early access.
Contact MyNew Technologies