Dariusz Kowalski
Polska wersja

FROM THE FIELD

LIVE

Five Bugs That Passed Every Test

Operational discipline is the layer no architecture diagram shows. Five production gotchas from a multi-agent QA system, and the lazy assumption behind each one.

  • #from-the-field
  • #operational-discipline
  • #production

FROM THE FIELD

LIVE

ADF Without Tears: The Full Pipeline and the Repo

The four-stage pipeline behind inline images in Jira: create, upload, resolve, embed. Plus the public AGPL repo you can clone and run against a mock Jira.

  • #jira
  • #adf
  • #ai-qa

FROM THE FIELD

LIVE

ADF Without Tears: The 303 Trick for Inline Images in Jira

Upload a screenshot to Jira and you get a gray External media box, not the picture. The fix is a 303 redirect and one fetch flag. Deterministic, no LLM.

  • #jira
  • #adf
  • #ai-qa

FROM THE FIELD

LIVE

From the Field #08, Part 3: Twenty Green Probes, Five Lessons, and the Pack

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.

  • #agentic-web
  • #a2a
  • #mcp

FROM THE FIELD

LIVE

From the Field #08, Part 2: Six Portals, Seventy Minutes, the Agent Layer

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.

  • #agentic-web
  • #a2a
  • #mcp

FROM THE FIELD

LIVE

From the Field #08, Part 1: Your Site Has a UX, Not an AX

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.

  • #agentic-web
  • #a2a
  • #agent-experience

FROM THE FIELD

LIVE

From the Field #07, Part 3: ROI and the Full Loop

What deterministic extraction returns, and where it goes. The generate-verify-fix loop as a commercial roadmap, not a shipped feature. AI does the dirty work, the engineer decides.

  • #figma
  • #codegen
  • #roi

FROM THE FIELD

LIVE

From the Field #07, Part 2: Two Tools, One Core

One deterministic extraction core, two public tools. figma-kit generates a component skeleton from real Figma values. qa-pack verifies a build against them, same RULE ZERO.

  • #figma
  • #design-tokens
  • #codegen

FROM THE FIELD

LIVE

From the Field #07, Part 1: The Data Layer Is Where It Breaks

Most design-to-code with AI is screenshot, paste, pray. That breaks at the data layer. Read the exact values from Figma, enforce RULE ZERO, and the model never gets to guess.

  • #figma
  • #design-tokens
  • #mcp

FROM THE FIELD

LIVE

From the Field #06, Part 3: What It Costs, What It Returns

The architecture decision is now an AI decision. The maintenance, scaling and ROI case for a deterministic test pattern, in the language of whoever signs the invoice.

  • #cdat
  • #testing
  • #roi

FROM THE FIELD

LIVE

From the Field #06, Part 2: Four Files, One Job Each

The four CDAT layers, the one-way dependency rule, and the same login test written two ways. Real code from a 9-system, 3000-test production pattern.

  • #cdat
  • #playwright
  • #typescript

FROM THE FIELD

LIVE

From the Field #06, Part 1: The God Object You Already Shipped

Before you point an AI agent at your test suite, the architecture has to hold on its own. Part 1 of a 3-part field report on the deterministic 4-layer CDAT pattern.

  • #cdat
  • #playwright
  • #testing

FROM THE FIELD

LIVE

Multi-page WCAG, Part 2: what site-wide compliance is actually worth

Your homepage passed its accessibility audit. That says almost nothing about whether your site is compliant - and since June 2025, that gap is a liability, not a nice-to-have.

  • #wcag
  • #accessibility
  • #compliance

FROM THE FIELD

LIVE

Multi-page WCAG, Part 1: the machine behind 5,816 to 7

5,816 findings across 35 pages. Four CSS commits. Seven false positives left. This is the machine behind it - the discovery layer, and why its default chain runs three strategies, not four.

  • #wcag
  • #accessibility
  • #multi-page

FROM THE FIELD

LIVE

Multi-page WCAG: 4 frameworks with full route-discovery, plus 4 recognised

Single-page audit = a Lighthouse extension. Multi-page = a different problem class. Route-discovery for 4 frameworks (Astro/Next/Vue/Nuxt), 4 more recognised with a warning.

  • #wcag
  • #accessibility
  • #multi-page

FROM THE FIELD

LIVE

Performance audit, Part 3: Where the method scales

Seven anti-patterns from a static pass, the line between a public AGPL toolkit and production work on three ecommerce platforms, and why /perf:fix never auto-fixes architecture.

  • #performance
  • #ai-tooling
  • #consulting

FROM THE FIELD

LIVE

Performance audit, Part 2: A deterministic floor you can trust

The hard part of a performance tool is not the AI - it is numbers you can trust without a human pasting them in. How I proved the measured floor holds before letting any AI speak.

  • #performance
  • #ai-tooling
  • #web-vitals

FROM THE FIELD

LIVE

Performance audit, Part 1: My own tool gave my portfolio a C

My own performance tool gave my portfolio a C while every Core Web Vital stayed green - LCP 991 ms, CLS zero, five areas graded A. Why one score lies and two axes tell the truth.

  • #performance
  • #ai-tooling
  • #web-vitals

FROM THE FIELD

LIVE

Context-First QA, Part 3: The Roadmap

Two days ago: the thesis. Yesterday: the map. Today: the calendar. Eight weeks. Four code drops. Four pitch-mode. Plus how to bring me in if the pattern fits - and why June 15 is the forcing function.

  • #ai-qa
  • #roadmap
  • #consulting

FROM THE FIELD

LIVE

Context-First QA, Part 2: The 10 Maps

Yesterday I showed the math. Today I show the map. Ten architectural layers that need to be deterministic before AI can do anything useful.

  • #ai-qa
  • #architecture
  • #three-layer

FROM THE FIELD

LIVE

Context-First QA, Part 1: The Thesis

I never let AI write my tests. I let it make decisions inside a deterministic harness. Here's the math, and why June 15 makes it visible.

  • #ai-qa
  • #manual-testing
  • #architecture

FROM THE FIELD

LIVE

When This Pays Off - And When Grep Is Already Enough

Day 3 of three. Should you adopt jarvis-brain at all? When graph-backed context pays off, when Grep is still enough, and what the V0.5 enterprise tier adds. From the field #02 Part 3.

  • #ai-tooling
  • #mcp
  • #context-engineering

FROM THE FIELD

LIVE

Architecture of the Indexing Engine

Day 2 of three. Inside jarvis-brain: code-to-graph extraction, the FTS5 camelCase trick, 5 MCP tools, and 50 questions worth of benchmark numbers. From the field #02 Part 2.

  • #ai-tooling
  • #mcp
  • #context-engineering

FROM THE FIELD

LIVE

Stop CC From Burning Tokens on Grep/Glob

Day 1 of three. Why Claude Code burns tokens on Grep/Glob in a 5-repo monorepo, and the architectural pivot that became jarvis-brain. From the field #02 Part 1.

  • #ai-tooling
  • #mcp
  • #context-engineering

FROM THE FIELD

LIVE

Scale Beyond the Distillate: F to A in 8 Commits, Plus What Pro Tier Actually Adds

Day 3 of a three-day live audit. F to A in 8 commits. Multiplicative token fix story. Pro tier walk-through: multi-runtime, auto-fix, niche specialists. From the field #01 Part 3.

  • #wcag
  • #accessibility
  • #tooling

FROM THE FIELD

LIVE

Triangulation: AI Specialists Across Three Audit Runs

Day 2 of a three-day live audit. V0.3 adds 5 AI specialists. Two runs find 16 unique findings, third finds zero. Convergence + dogfooding bug fix. From the field #01 Part 2.

  • #wcag
  • #accessibility
  • #tooling

FROM THE FIELD

LIVE

When axe-core Isn't Enough: Auditing My Own Portfolio with V0.2 Public

Day 1 of a three-day live audit. V0.2 public WCAG toolkit catches 3 real findings on portfolio.sdet.it. Static + dynamic baseline, no AI yet. From the field #01 Part 1.