Figure 1. NeuroCW at a glance. All CER numbers are from synthetic test sets. ▶ New: watch the first on-air test of the NeuroCW test build NeuroCW is an end-to-end neural CW decoder: audio in, text out. One Conformer-CTC model copies English, Korean (Hangul) and Japanese (Wabun). It was trained for about two weeks on an NVIDIA DGX Spark, exported to ONNX, and a Windows release is planned for early November . This post walks through how it was built, mostly in pictures. 1. The problem Classic decoders do fine on clean signals. They struggle once QSB, QRM and a human fist get involved. NeuroCW was trained on the bottom case. Figure 2. The same transmission, clean (top) and after the synthetic HF channel (bottom). Figure 3. End-to-end pipeline: generator → spectrogram → Conformer → CTC text. 2. Synthetic data, generated on the fly No dataset sits on disk. Every batch is generated fresh and is deterministic per seed. The text, the operator's fist and the radio channe...
TL;DR: Open WebUI builds below v0.1.124 carry CVE-2026-44566 ( GHSA-9pgh-j74g-qj6m ) — an arbitrary file upload / path traversal weakness. In a controlled lab demo against a disposable v0.1.120 container, a traversal write reached outside the uploads directory and overwrote /etc/passwd inside the Docker container, injecting a visible hacked account marker. This is not a reverse shell — it is a file-write-to-anywhere impact. Patch now — upgrade to a fixed release, rebuild images, and lock down upload endpoints. ⚠ Security notice: This post is for defensive awareness only . The demo was run in an isolated lab against a disposable target I own and control. This article contains no payloads, no exploit code, and no tooling for download , and no step-by-step reproduction. Do not test systems you do not own or lack explicit permission to assess. Demo intro — Open WebUI CVE-2026-44566 lab thumbnail Watch the lab demo I recorded a short proof-of-concept walkthrough (English ...