Research for Practical Web Technology

We are investigating how Agent Skills — curated Markdown knowledge packages — equip AI coding agents with genuine TYPO3 knowledge. The core question: how do you establish trust in an ecosystem where even the prose of a manual can become an attack vector?

What started as an internal strategy has become a dissertation-style monograph: specification catalogue, Git-first architecture, governance, legal framework, and an adversarial risk analysis – openly readable and with clearly defined boundaries.

Read the research paper · Ten-minute summary

We develop quality analysis tools for SME websites that go beyond superficial metrics. Our system correlates performance data with actual user behaviour and identifies concrete optimisation potential – from loading speed to conversion rate.

The analysis is based on measurable quality criteria: Core Web Vitals, UX patterns, and accessibility standards. We deliver the results as a prioritised roadmap with concrete recommendations – no generic reporting, but directly actionable optimisation steps.

The goal is to establish best-practice standards for regional SMEs. In parallel, we develop automated toolchains for continuous quality assurance – ensuring website optimisation is not a one-off project, but an ongoing part of operations.

We are happy to share the detailed study results with interested parties: office@webconsulting.at

Our automated analysis tool systematically analyses websites for technical performance, SEO quality, and content effectiveness. The analysis starts by entering a URL and delivers structured reports on framework configuration, loading times, and optimisation potential.

The system checks SEO factors, Core Web Vitals, WCAG compliance, and content quality. AI-driven algorithms identify vulnerabilities and generate prioritised lists of measures – sorted by effort and expected impact.

The AI component is continuously trained: through user feedback and the analysis of current web standards, we improve the precision and relevance of recommendations. The goal is a tool that does not just analyse, but advises strategically.

We are developing a modern backend prototype for TYPO3 that simplifies content management workflows. The focus is on intuitive navigation, faster load times, and a consistent user experience – factors that directly influence the productivity of editorial teams.

The technical foundation is a React-based tech stack with Next.js components. This architecture enables responsive behaviour, modular design, and a clear separation between backend logic and frontend interface – ideal for future enhancements such as AI-driven content suggestions.

The architecture is designed for integration: we plan to connect it to the TYPO3 v14 REST API. The modular structure already allows adjustments for specific use cases – from multilingual editorial systems to complex approval workflows.

TYPO3 updates are time-consuming – our automated updater reduces the time required by up to 90%. The system fully handles project detection, Composer updates, and Git integration. Development teams gain time for strategic tasks, while the risk of faulty updates is reduced.

Visual regression tests ensure stability: before every deployment, the system creates baseline screenshots using Playwright. After the rollout (via Deployer), new screenshots are automatically compared. Discrepancies are reported immediately – before users are affected.

The intelligent project detection analyses DDEV configuration, Composer dependencies, and update history in a structured manner – without AI guesswork, with deterministic reliability. Transparent documentation and ticket system integration complete the workflow. The result: updates become predictable, traceable, and secure.