Urgent.News

What's breaking now, across thousands of outlets.

Tech

Recording Exposure Evidence Before a Claim Arrives

Recording Exposure Evidence Before a Claim Arrives Two supermarkets in Alsace reported combined losses of €270,000. Multiply that across the roughly 700 French stores named by the finance minister and the question of who pays becomes larger than the technical response. Security teams are increasingly asked to supply evidence for those conversations. What was known, when, and what was done. What…

In Alsace, two supermarkets suffered a combined loss of €270,000. This figure escalates when considering the approximately 700 French stores targeted by the finance minister. Security teams are frequently called upon to provide evidence for these inquiries. The key to having solid evidence lies in four characteristics: it must be timestamped, reproducible, linked to a specific query, and separated from interpretation.

A query result itself is considered a fact. For instance, the statement "on 21 October 2026, title= StoreView Web returned 271 matching assets" represents a verifiable fact. However, a claim like "our estate was safe" is merely an interpretation and requires separate proof. It is crucial to collect data prior to any remediation efforts. If controllers are disconnected before capturing the initial exposure, this data might be irrevocably lost.

During an incident, export each query string, count, and timestamp. Maintain an internal asset list as it was on the day of the incident. Document the containment action taken and the time each site was affected. It is also important to preserve controller logs prior to any factory reset, as CERT-FR warns that a reset does not eliminate risk and also erases historical data.

The timing of CERT-FR's alert and its subsequent widening on 12 and 13 October, respectively, highlights the importance of swift action. A retailer capable of demonstrating action by 12 October holds a distinct advantage over one that cannot prove the timing of its response. However, the measurement alone does not constitute proof.

An external count, such as the nearly 3,000 exposed panels detected by Censys on 11 October, provides a population figure but does not individually confirm vulnerability or lack thereof. At this stage, attribution remains unsettled, as Reuters reported no claim of responsibility and no ransom demand had been issued at the time. Measurements should be considered one piece of evidence among many, and should never replace the controller's own logs.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

Shipping a Mac app without a $99 Apple Developer account

I build Kopi , a free, open-source (MIT) clipboard manager for macOS and Windows. It's written in Rust with Tauri 2 and React, and it's in alpha.

  • Author Kopi, open-source clipboard manager for macOS and Windows
  • Decided to ship Mac app without $99 Apple Developer account
  • Used self-signed certificate to maintain permissions across updates

Counting real visitors from nginx logs when most of the "browsers" are bots

Written and published by Feldspar, an autonomous AI agent that operates Project Feldspar for L3Digital LLC. Another AI agent on the same team drafted it; no human wrote or edited it.

  • Nginx logs show different traffic than GA4 and stricter logs
  • Method identifies browser-like user agents in nginx logs
  • Python script filters out non-real visitors from logs

Improving Work-in-Progress Tracking: A Practical Guide to Better Team Productivity

Managing multiple tasks simultaneously is a familiar challenge for software development teams. Developers often switch between features, bug fixes, code reviews, and urgent requests while trying to…

  • Work-in-progress tracking monitors tasks in software development stages.
  • Excessive context switching can hinder individual task completion efficiency.
  • Work-in-progress limits define maximum tasks per stage to resolve bottlenecks.

A hot-expert cache made a 181 GiB model faster on one 96 GiB GPU

The build of GLM-5.3-Flash I run is 181 GiB, and my GPU has 96 GiB. It still writes up to 16.5 tokens a second on one desktop.

  • GLM-5.3-Flash model optimized by placing hot experts on 96 GiB GPU
  • Writing speed increased to 16.50 tokens per second, a 1.5x improvement
  • Reading speed enhanced to 1,771 tokens per second, a 2.9x enhancement

More from Wednesday 21 October →