Notes

Notes from engineering.

At the top are articles that grew out of our own work: building the coding agent, the Sentinel validation layer and projects with a regulatory angle. Below follows a fundamentals section for reference. Every article appears in German and English.

AI Agents

Context that does not overflow

Our coding agent manages its context window in two tiers. The run logs showed placeholders that invited repetition, and one overflow.

10 October 2026 · 7 min read
Security

Three KQL errors the parser waves through

Three Sentinel rules are valid KQL and stay silent in operation. Our check reports all three as errors. With the findings quoted verbatim.

10 October 2026 · 8 min read
Robotics

Simulate first, then build

Our humanoid robot exists as a design and a simulation. The simulation follows three rules; for one of them we withdrew our best results.

10 October 2026 · 5 min read
AI Agents

When a coding agent may say done

A coding agent reported done without ever running its code. Since then the last run decides, and plan steps need a green gate command.

10 October 2026 · 7 min read
AI Agents

Running Coding Agents Safely

A prompt alone does not secure a coding agent with shell access. We secure our own agent with a permission engine, a verified sandbox and undo.

9 October 2026 · 7 min read
Security

Why 13% of SIEM Rules Never Fire

CardinalOps found that 13% of rules in production SIEMs never fire. Sentinel rules can be checked deterministically against Microsoft’s KQL parser.

17 August 2026 · 9 min read
NIS2

NIS2 Incident Reporting in Practice

NIS2 requires an early warning within 24 hours of awareness. The hurdle is detecting the incident. What Article 23 requires and what you build for it.

11 August 2026 · 8 min read
Security

Detection Engineering with LLMs

An LLM writes a detection query in seconds. Whether it is correct only shows in a generate-validate-refine loop that runs it against the schema and sample data.

10 August 2026 · 8 min read
RAG

Long Context vs. RAG: Which Do You Need?

Million-token context windows did not replace RAG. Long context stays affordable only when the prompt prefix is identical for every query.

6 August 2026 · 9 min read
Contracts

Acceptance Criteria for Non-Deterministic AI

An AI system returns a different number on every measurement. Acceptance becomes testable when the contract owes a threshold on a frozen data set.

15 July 2026 · 8 min read
Co-Determination

AI and the Works Council

Almost any AI tool your staff use triggers co-determination under § 87 BetrVG. The four provisions that apply and what a works council agreement covers.

14 July 2026 · 9 min read
GDPR

AI and GDPR in Practice

How to build AI on personal data in line with the GDPR: which duties apply, where projects fail, and which architecture anchors the duties in the system.

5 July 2026 · 9 min read
On-Premise

On-Premise LLMs: A Practical Hardware Guide

Hardware for local LLMs: choose a model size, estimate GPU memory (VRAM), save with quantization, and know when on-premise beats the cloud.

4 July 2026 · 9 min read
Fundamentals

For reference

Introductions to the terms and methods the articles above build on.