Sauce Labs launches AI agent for natural language test authoring

Sauce Labs has launched Sauce AI for Test Authoring, an AI agent that translates natural language business intent into executable, self-improving test suites. The platform targets the growing gap between AI-accelerated code production and slower test validation, claiming up to 90% faster test creation and 41% faster issue diagnosis. It leverages a dataset from 8.7 billion real-world test runs as a competitive differentiator. The tool enables non-engineers like product managers and QA analysts to contribute to test automation without coding. Competitors in the space include TestMu AI (KaneAI), mabl, Testsigma, Katalon, and Testim, all offering similar natural language-driven or AI-assisted test generation capabilities.

Red Hat turns Storybook into a behavioral verification engine for 959 CI specs

Red Hat's team transformed Storybook from a component playground into a full behavioral verification engine for their Hybrid Cloud Console access management UI. Using Mock Service Worker (MSW) to intercept HTTP at the network boundary, they built a system of typed handler factories, seed constants, shared interaction helpers, and step-isolated journey stories. 959 specifications run in CI on every change, catching regressions across large refactors. Beyond code quality, the hosted Storybook environment became a shared artifact for design reviews, product walkthroughs, and stakeholder demos — replacing staging and screen recordings with live, accurate component behavior.

Grafana Cloud k6 adds secrets management for performance test scripts

Grafana Cloud k6 now offers secrets management in public preview, allowing teams to securely store API keys, tokens, and credentials centrally rather than hardcoding them in test scripts. Secrets are injected at runtime via a dedicated `k6/secrets` module, are write-only in the UI (never displayed after creation), and are automatically redacted if accidentally logged during test execution. The feature supports full lifecycle management (create, edit, delete) with labels and descriptions, and is also generally available in Grafana Cloud Synthetic Monitoring.

DoorDash migrates iOS shared codebase from XCTest to Swift Testing with 60% faster CI

DoorDash migrated its iOS shared codebase from XCTest to Swift Testing, achieving ~60% faster CI test steps and ~40% faster overall builds. The team built a migration environment using Cursor (AI coding assistant), SweetPad (VS Code/Cursor extension for Apple platform development), and a custom MCP server to automate mechanical code transformations. A Cursor rule handled API replacements (XCTAssert* → #expect, XCTestCase → struct + @Suite, etc.), while Bazel and SweetPad provided a reliability gate requiring 10 consecutive clean test runs per package. The migration exposed previously hidden flaky tests caused by XCTest's sequential execution model, which Swift Testing's parallel-by-default approach revealed. Results include CI cost savings of several thousand dollars annually, multiple engineer-days per week reclaimed, and 90% coverage of shared iOS modules migrated with no regressions.

Uber migrates 75,000 test classes from JUnit 4 to JUnit 5 using automated code transformation

Uber engineers migrated over 75,000 test classes and 1.25 million lines of code from JUnit 4 to JUnit 5 across their Java monorepo using automated code transformation. The team used OpenRewrite for deterministic source transformations and an internal orchestration system called Shepherd to apply changes across thousands of Bazel targets in parallel. Key challenges included Bazel's lack of native JUnit 5 support (solved via a unified JUnit Platform execution model with Vintage and Jupiter engines) and Uber-specific test patterns requiring custom transformation recipes. Generative AI was evaluated but rejected for inconsistent results. The iterative rollout validated changes through CI pipelines, achieving large-scale migration with minimal manual intervention. The approach is now being extended to Spring Boot 3 and other library migrations.

KubeStellar reaches 81% AI PR acceptance rate using structured test coverage and feedback loops

A solo developer building KubeStellar Console — a multi-cluster Kubernetes dashboard — shares how AI coding agents initially caused chaos before a structured 'AI Codebase Maturity Model' turned things around. The model has five levels: Instructed (externalizing preferences in instruction files), Measured (91% test coverage across 32 nightly suites with strict determinism), Adaptive (automating only after measurement, with dynamic category weighting), Self-Sustaining (codebase encodes enough judgment to operate autonomously), and a prompting habit of asking 'why' instead of 'what.' The result: 81% PR acceptance over 82 days, 30-minute bug-to-fix cycles, and a system that triages, fixes, and explains design decisions without human intervention. The core thesis is that AI leverage comes from the feedback infrastructure surrounding the model, not the model itself.

BugBash'26 day one covers distributed systems testing, formal verification and property-based testing

Conference notes from BugBash'26 morning sessions covering several talks: Peter Alvaro's 20-year retrospective on distributed systems research (CALM theorem, lineage-driven fault injection, metastability simulation, deterministic testing); Deb Chachra's infrastructure philosophy talk drawing parallels between physical and digital infrastructure; and lightning talks on kernel bug patterns and LLM implications, formal verification with Dafny in web dev workflows, and property-based testing for web apps using the Bombadil project from Antithesis.

RV Tech deploys AI agent to generate tests for safety-critical propulsion code 10-15x faster

<p>Rivian and Volkswagen Group&#39;s joint venture put Devin to work on testing and ticket triage across a software platform that will power up to 30M vehicles.

Devin handles autonomous ticket triage in Slack and generates tests on safety-critical propulsion code 10-15x faster than manual authoring, freeing RV Tech&#39;s engineers to spend more time on new features.</p>

The joint venture between RV Tech (formed by the partnership of the two automakers) deployed the AI coding agent, to automate ticket triage in slack and generate tests for safety-critical propulsion code. The result is a 10-15x speedup in test authoring, freeing engineers to focus on new features across a software platform targeting up to 30 million vehicles.