LOAD "*",8,1
Context before LLM.
Portfolio of Dariusz Kowalski. I build multi-agent QA systems, AI pipelines and developer platforms. CDAT Pattern, Jarvis Platform, open-source distillates.
Components · Data · Actions · Tests. 4-layer Playwright architecture battle-tested across 9 production systems over 2 years.
ECOSYSTEMPrivate multi-agent QA platform. 34K LOC TypeScript, 9 microservices, 15 production pipelines. Ask for demo.
THESISAI in QA does not start with "write me a test". It starts with deterministic, pre-processed context. LLM comes second.
PIPELINE7 agents parallel: keyboard, forms, modals, contrast, ARIA, semantic HTML, e-commerce. WCAG 2.2 AA full audit on live production.
PIPELINE5 agents parallel: bundle, Vue/React runtime, API calls, SSR/hydration, assets. 7h with AI vs 16h billable vs team-week classical.
PIPELINEDeterministic Figma pipeline. CSS token mapping, pixel diff via odiff, codegen spec. Design tokens at data layer, LLM at logic only.
FROM THE FIELD
A 200 is not proof an agent can read you. Twenty smoke probes across six portals, five lessons from the build, and a clone-and-run pack so you can ship the same layer.
FROM THE FIELD
A custom A2A Agent Card and MCP discovery manifest per portal, one Astro pattern that serves two languages from one source, and the portal that got bearer-gated discovery on purpose.
FROM THE FIELD
Your site was built for a human to experience. An AI agent that lands on it sees nothing it can act on. That gap has a name now: Agent Experience, and almost nobody ships it.
Production-ready manual QA workflow extracted from Jarvis. Context-first pipeline: Figma MCP + Jira webhook + Playwright CLI + Claude Agent SDK. Scale: 100-200 tasks in 2-3 days vs team-week classical.
Components-Data-Actions-Tests - 4-layer architectural pattern for Playwright + TypeScript. Alternative to Page Object Model. Battle-tested across 9 production systems over 2 years.
Public AGPL-3.0 distillate of multi-agent WCAG audit pipeline. 5 AI specialists reading source via Read/Grep/Glob, plus static TypeScript analyzer and Playwright + axe-core dynamic testing. A-F grading. Case study in From the Field series #01.
Local-first persistent RAG for personal Markdown corpus. Qdrant + MLX + FastAPI + FastMCP 3.0. 12 MCP tools, 213 tests, source-available. Replaces copy-paste of context across Claude Desktop / Code / OpenCode chats.
Open-source MCP server fixing Claude Code skill bloat. Two-Tier Discovery: ~1k token mini-index always preloaded, full SKILL.md loaded on demand. 68% token reduction at 60 skills, roughly flat at 500. Hybrid retrieval (BM25 + dense), trust tiers, 100% local Apple Silicon stack via MLX (Qwen3-Embedding-8B + Qwen3-Coder-30B rewriter/reranker). No Ollama, no HTTP, no network.
Performance audit for Nuxt3/Vue3. Deterministic floor: Lighthouse median-of-5 + Core Web Vitals via Playwright + trace + bundle. 5 AI specialists for what the floor does not catch. AGPL-3.0.