Flaky-test triage, bug reproduction, defect management, load testing, resilience drills, and shift-left and shift-right checks.
Bug reproduction and triage: well-formed bug reports, minimal failing tests built from issue reports, and CI failure classification before a fix starts. Also classifies production escapes by root gap. Agents (bug-report-from-recording, bug-repro-builder, defect-pipeline-runner) turn Playwright traces and HAR files into bug reports, build the repro test, and run the weekly defect review that clusters the backlog and narrates trends.
Defect management: severity vs priority classification, bug tracker workflows on Jira, Linear, GitHub Issues and Azure DevOps, and fix confirmation. Agents (ci-defect-filer, defect-fix-verifier) turn CI failures into deduplicated tracker bugs and re-run a defect's reproduction once a fix is claimed. Distinct from qa-bug-repro, which covers reproduction, report authoring, and the weekly defect review.
Flaky test triage: find the condition a flake depends on, quarantine it with an expiry date, fix known flake patterns, and track flake rates. Agents (e2e-flake-bisector, regression-bisector) rerun a test under varied conditions with a shared-state isolation stage, and run git bisect to find the commit that introduced a regression, with a perf-measurement mode.
Load and performance testing with k6 and JMeter, Lighthouse CI for Web Vitals, SLO-driven load plans, and performance budget gates. Covers tool choice and load profiles (with Gatling and Locust references), per-interaction INP budgets, flame-graph analysis, and database query-plan analysis.
Chaos engineering and resilience drills: chaos experiment design, Chaos Mesh and Toxiproxy fault injection, disaster recovery drills, error-budget gates. Covers steady-state hypotheses with LitmusChaos and Gremlin routing, a drill run protocol with pre-flight and abort gates, combined HTTP and TCP fault injection, backup restore and RTO measurement, and burn-rate alerting with MTTR / MTBF / MTTD / MTTA incident metrics.
Shift-left QA: testability reviews of specs before implementation, non-functional requirements turned into testable thresholds, and STRIDE threat modeling. Acceptance-criteria extraction and Gherkin coverage mapping live in qa-bdd; data-contract extraction lives in qa-data-quality.
Shift-right production QA: synthetic monitors, canary validation against baseline metrics, and release runbooks. Follows the ISTQB definition of shift right (testing a system continuously in production) and covers RUM coverage gaps and multi-team cutovers. An agent (observability-to-test) turns production alerts and errors into regression tests.
