Welcome to Issue #2!
Cypress 15.16 just dropped. Appium 3 migration questions are picking up on every QA channel I follow. And the “AI will replace testers” conversation has shifted in an interesting direction this week. More on that below.
Back next week with more signal and less noise. 🙂
NEWS
Cypress 15.16 Out This Week, Studio Beta Now On for Everyone
The big news buried in the 15.x cycle: experimentalStudio flag is gone. Studio Beta is now on by default for all users, no config needed. The team also renamed Cypress.SelectorPlayground to Cypress.ElementSelector, added Vite 7 and Angular 20 support, and dropped Node 18 and 23. If you’re still on those Node versions in CI, plan the upgrade before this bites you.
docs.cypress.io 🔗
Appium 3 Is Here and the Migration Is Simpler Than You Think
Appium 3 (now at 3.2.2) is less of a revolution and more of a long-overdue cleanup. Express 4 to 5, Node 20.19 minimum, deprecated endpoints removed, feature flags now require driver-name prefixes like uiautomator2:adb_shell. The official migration guide is genuinely short. Most teams on Appium 2.x will get through it in a day. The bigger ask is upgrading your Node version in CI if you haven’t already.
appium.io 🔗
The AI Testing Platform Race Is Heating Up
Every major testing vendor suddenly has an AI story. From autonomous agents to self-healing tests, the race to become the go-to AI testing platform is accelerating. The marketing is getting louder. The real question is which tools can actually reduce maintenance and improve test quality at scale.
softwaresuggest.com 🔗
AUTOMATION
Testcontainers on GitHub Actions: Real Dependencies, Zero Setup Overhead
Docker’s guide walks through running Testcontainers-based tests on GitHub Actions using Testcontainers Cloud to offload container management from the runner. The core appeal: your integration tests hit a real PostgreSQL, Redis, or Kafka instance, not a mock. Ubuntu runners have Docker pre-installed, so the baseline setup is straightforward. The Cloud token approach is what makes it scale cleanly in parallel runs.
docker.com 🔗
Agentic AI Testing: What It Actually Means in Practice
Autify draws a useful distinction worth keeping: AI testing is a broad label covering generation, self-healing, and prediction. Agentic AI testing is specific — autonomous agents that plan, reason, and execute without scripts or selectors. The practical claim worth noting: because agents use visual recognition and semantic understanding rather than DOM selectors, they’re supposed to adapt when UI changes instead of breaking. That’s the theory. I’d want to see that hold up on a real design system at scale before betting a regression suite on it.
autify.com 🔗
Playwright Network Interception: Every Request Passes Through Your Handler First
Clear write-up on how Playwright’s route handlers work: intercept at page or context level, then fulfill with mock, continue to real server, or abort. The part most teams underuse is context-level mocking, which applies the same rule across every page in a test session. Useful for blocking analytics calls globally without repeating the route in every test.
oneuptime.com 🔗
TOOLS & GITHUB
cypress-io/github-action v7.3.0: Adds expose Input for Cypress.expose() API
Kudos to the Cypress team for keeping this action well maintained. v7.3.0 adds an expose input to surface the Cypress.expose() API, v7.2.0 dropped Node 20 support (now requires Node 22). Worth checking your workflow’s node-version setting if you’re pinned.
github.com 🔗
testcontainers — Now Available for Node, Python, Go, Rust, Elixir, and .NET
The Testcontainers org has grown well beyond Java. If you’re running integration tests in any of these languages and still setting up shared test databases, worth a look. The Node library in particular is clean and pairs well with Playwright for full-stack test isolation.
github.com 🔗
github-action-playwright-flaky-test-analyzer: AI-Powered Flake Diagnosis in CI
Docker Agent that identifies and diagnoses flaky Playwright tests with browser-specific insights. Processes JUnit XML, analyzes traces and screenshots, and provides Playwright-specific fixes: proper waits, stable selectors, network mocking suggestions. Runs on Claude Sonnet 4. Worth trying on a suite that’s been accumulating unexplained retries.
hub.docker.com 🔗
COMMUNITY INSIGHT
QA to SDET in 2026: The Role Is Shifting Faster Than Most Job Descriptions Reflect
Quash’s April write-up captures something I hear a lot in hiring conversations right now: the SDET title is being stretched to mean something genuinely different in 2026. The teams moving fastest aren’t just adding AI tools on top of existing workflows, they’re redesigning the role around tool-calling, agent orchestration, and test strategy rather than script maintenance. The line “future-proof career: master tool-calling, evolve from QA to AI Automation Architect” is a bit dramatic but the direction is right.
quashbugs.com 🔗
88% of Developers Aren’t Confident Deploying AI-Generated Code. Who’s Testing It?
TestDevLab’s write-up cites the Veracode stat that AI-assisted code development correlates with measurable increases in security vulnerabilities. Gartner says 33% of enterprise apps will include agentic AI by 2028. The question nobody is answering cleanly: if the code is AI-generated and the tests are AI-generated, what does a failing test actually tell you? That’s not a rhetorical question. It’s the real design problem for the next generation of QA infrastructure.
testdevlab.com 🔗
PRACTICAL TIP
Pin Your Testcontainers Image Tags. Always
When using Testcontainers, never pull :latest. Always pin to an exact image tag like postgres:16.2-alpine. Latest changes silently across environments and CI runs, which produces the exact flakiness Testcontainers is supposed to eliminate. A test that passes on your machine and fails on the runner because the image updated overnight is infuriating to debug. One line in your container definition, saves hours of confusion.
oneuptime.com 🔗
VIDEOS
What’s New in Appium 3.0: Changes You Must Know
Quick and practical walkthrough of everything that changed in Appium 3. Good first watch before you touch the migration guide. Covers the breaking changes clearly without padding.
youtube.com🔗
Agentic Mobile Testing That Fixes Its Own Tests — Maestro MCP
Joe Colantonio talks to Maestro co-founder Leland Takamine about AI agents that don’t just write mobile tests but validate and debug their own output automatically. Live demo included. If you’re evaluating agentic tooling for mobile specifically, this is more grounded than most vendor content on the topic. Maestro is already in use at Microsoft, Meta, and Amazon.
testguild.com 🔗
This Week’s Discussion
My pick: “Open Question”
Every SDET job posting in 2026 says something different.
Some want Playwright engineers. Some want AI agent builders.
Some still list manual test execution under the same title.
The role is either genuinely evolving or just getting
relabeled depending on who you ask.
What does your SDET role actually look like day to day?
→ Join the discussion on Reddit
And Finally…
Reviewed a team’s Appium 3 migration plan this week. Solid work, but they’d missed the feature flag prefix change entirely. Three hours of “why is adb_shell not working” could have been avoided by reading the two-page migration guide. Always read the migration guide. 😅
