Rethinking how we build software in the era of AI

Latest Writing

All articles →

Five Gaps in Agent Skills and What Belongs in the Context Layer

The Agent Skills specification defines a context layer for AI agents: instructions, domain rules, and supporting files that travel with the codebase and shape agent behaviour. The format succeeded by staying minimal. Five gaps around variables, triggers, scoping, composition, and workflow execution are now generating proposals that would close them by adding execution logic to the format. Each gap has a cleaner answer — one that keeps the context layer honest and the orchestrator responsible.

Agentic Coding Skills Developer Experience Architecture
13 min read

The Agent is not the Pipeline: Spec-Kit Workflows and the Enforcement Layer

Agents are probabilistic by design. Hard workflow guarantees — phase gates, enforced sequencing, human checkpoints — cannot be achieved by giving an agent better instructions. They require a runtime the agent cannot override. Spec-Kit's workflow engine provides one.

Agentic Coding Spec Kit Architecture Workflows

Selected Work

BriefCheck App

Document Analysis & Agentic Assistant

AI-powered letter analysis that helps you understand bureaucratic correspondence from authorities, digitizes documents, and delivers clear action recommendations: fast, secure, and simple.

Visit BriefCheck
AI Document Analysis NLP

AgentMux

Software Engineering Pipeline

Orchestrates a full PM → Architect → Code → Review pipeline across Claude, Codex, and Gemini, injected via tmux, not the API. Reuses your existing subscriptions. No per-token costs.

Python tmux Multi-Agent

Finance App

Personal Project

Modern personal finance tracking with wealth management, fixed cost analysis, and financial forecasting. Fintech-inspired dark theme.

Vue.js Go

Wondrax Platform

Infrastructure

Self-hosted infrastructure automating containerised application deployment on a single VM. Automated pipelines, shared database; low-maintenance and production-ready.

Infrastructure Docker CI/CD

Faststring

Master's Thesis

An algorithm that exchanges arbitrary types in Java bytecode at runtime. Identifies safe boundaries to seamlessly convert legacy data structures without recompilation.

Java Bytecode

About

Markus Wondrak

Software engineering enthusiast with a clear mission: making life easier for developers and producing good-quality software.

After studying computer science, I spent 9 years in consulting before becoming a Domain Architect at a major German bank, supporting their digital transformation.

Currently deep in agentic coding and the toolchain around it: exploring how these tools integrate into software development and reshape the industry.

GitHub LinkedIn Wondrax Solutions