🧩 The Edge Case | AI QA Workflows Get Practical ✅

Sponsor Spotlight • Level 1

Partner with QA Testing Jobs to feature your tools, platforms, or services to thousands of testers every week.

Hi there,

AI testing moved from “write more tests” toward agent-governed quality loops: better browser automation primitives, AI-review instructions, usage/cost telemetry, and production-trace-driven evals.

This Week’s Highlights

1. Playwright v1.61 adds passkey/WebAuthn and storage testing primitives

Sources: Release notes, Credentials API

Playwright v1.61 introduces stronger support for passkey and WebAuthn automation, plus improved storage testing capabilities.

Why it matters:
Teams can now automate passkey login and registration flows without relying on physical hardware security keys. This makes it easier to test modern authentication journeys, seed or capture credentials, manipulate local/session storage, and retain richer videos or traces on retry.

Useful for: auth-heavy regression suites, flaky login flows, and teams modernising identity testing.

2. Selenium 4.45.0 released

Sources: GitHub release, PyPI, Maven

Selenium 4.45.0 landed with continued improvements across the Selenium ecosystem, including Grid, WebSocket, language bindings, and BiDi-related work.

Why it matters:
If your organisation still relies on Selenium Grid, this release is worth evaluating for reliability improvements, especially around intermittent session, proxy, or driver-related issues.

Useful for: enterprise regression suites, cross-browser testing, legacy applications, and teams with existing Selenium infrastructure.

3. GitHub Copilot Agent Finder is now available

Source: GitHub changelog

GitHub Copilot can now discover approved MCP servers, tools, agents, and skills on demand rather than requiring everything to be preloaded into context.

Why it matters:
This points toward governed internal agent catalogues. QA and delivery teams could maintain approved tools such as a “test data agent”, “API contract checker”, “Playwright helper”, or “release readiness assistant”.

Useful for: organisations introducing AI agents while still needing governance, access control, and standardised tooling.

4. Copilot Code Review now supports repo-level AGENTS.md

Source: GitHub changelog

GitHub Copilot Code Review can now use repository-level AGENTS.md files to understand project-specific instructions.

Why it matters:
Teams can encode review expectations once, including test coverage rules, flaky-test policy, selector strategy, accessibility checks, API backward-compatibility rules, and expectations such as “update Playwright tests when routes change”.

Useful for: improving AI-assisted code review consistency and embedding QA expectations directly into repositories.

5. Copilot usage metrics API now reports per-user AI credits

Source: GitHub changelog

GitHub’s Copilot usage metrics API now includes AI credits consumed per user.

Why it matters:
Engineering and QA leaders can start correlating AI-tool cost with adoption, productivity, and delivery outcomes. This is especially useful when rolling out AI coding agents to testers and developers and needing budget guardrails.

Useful for: AI adoption reporting, usage governance, budget management, and rollout planning.

Suggested Experiments This Week

1. Add an AGENTS.md file to one repository

Create a lightweight repo-level instruction file with QA-specific review rules, such as:

• Required tests for bug fixes
• Playwright locator conventions
• API contract expectations
• Accessibility checks
• Flaky-test handling rules
• Expectations for updating tests when application routes or flows change

Then request a Copilot review on a draft PR and compare whether the feedback becomes more relevant and consistent.

Takeaway

The main trend this week is not just better test tooling, but more governed and measurable AI-assisted quality workflows. Testing teams should start thinking about AI agents the same way they think about automation frameworks: with standards, review rules, traceability, regression checks, and clear ownership.

Yours in testing,
Farid Vaswani

Community Partner

Have a QA community, course, or event to promote? Reply to reserve this feature placement.