🧩 The Edge Case | Passkeys, Cypress stability, and safer AI testing ✅

Sponsor Spotlight • Level 1

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

Hi there,

This week’s useful testing updates were mostly about control: making tricky browser flows easier to automate, reducing noisy test-runner failures, and putting better limits around AI-assisted development.

For Software Testers, the theme is simple: the next productivity gain is not just “more AI” — it is better test evidence, clearer gates, and safer automation around the work.

This week’s useful testing and automation updates

• Playwright 1.61 adds virtual WebAuthn passkeys and Web Storage APIs — Playwright can now register and answer passkey flows without real hardware keys, and tests can read/write local and session storage directly. This is useful for teams testing modern login, account recovery, and authenticated journeys that were previously hard to cover reliably in CI.

• Cypress 15.18.0 improves stability around repeated uncaught exceptions — Cypress now collapses repeated identical uncaught exceptions instead of letting them exhaust renderer memory, and adds a removeSRIAttributes option for apps where Cypress rewriting conflicts with Subresource Integrity. Worth reviewing if your E2E suite has unexplained browser crashes or issues with lazily loaded assets.

• Claude API rate limits were raised and usage tiers simplified — Anthropic says Sonnet and Haiku API limits now match Opus at each tier, with tiers consolidated into Start, Build, and Scale. For testing teams experimenting with AI-generated test ideas, triage summaries, or eval runs, this may make small internal pilots easier to run without hitting limits so quickly.

• Claude SDKs added support for newer code execution tooling — the Python, TypeScript, Go, Java, Ruby, PHP, and C# SDKs now support a code execution tool version with REPL state persistence and programmatic tool calling support. If you are testing AI agents, this is a reminder to evaluate not only final answers, but also tool calls, intermediate state, and execution boundaries.

• GitHub Code Quality moves to GA on July 20 — GitHub is positioning Code Quality around maintainability, reliability, coverage enforcement, quality gates, dashboards, and PR-blocking rulesets. QA and engineering leads should review whether coverage or reliability thresholds belong in their merge rules, and whether any billing or repository enablement changes need planning.

A few experiments to try this week

1. Add one Playwright spike for a passkey login or WebAuthn registration flow, then check whether it can run reliably in CI without manual hardware.
2. Review your Cypress failure history for repeated uncaught exceptions, browser crashes, or SRI-related asset failures, then trial Cypress 15.18.0 in a sandbox branch.
3. Create a small AI-testing eval sheet with 20 examples: generated test cases, expected risks, tool calls used, and a human pass/fail judgement.

Browse QA roles when you are ready: https://qatestingjobs.com/jobs

Yours in testing,

Farid Vaswani

Community Partner

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