Ismail 3000 ยท QA x AI articlesIsmail 3000, the reference for AI-augmented QA: articles, downloadable kits, AI testing agents and support for startups. By Ismail Haiber, Full-Stack QA Engineer in Paris.https://ismail3000.fr/enPHP2Shell: what QA (and AI) could have changedhttps://ismail3000.fr/en/articles/failles-php-ce-que-la-qa-aurait-pu-changer/https://ismail3000.fr/en/articles/failles-php-ce-que-la-qa-aurait-pu-changer/Two critical PHP flaws, spotted as early as 2024 but never fixed. This is not only a security story: it is the cost of QA that checks features and neglects non-functional quality. And a lesson in what AI really changes.Sat, 31 Oct 2026 17:09:00 GMTAI & QAsecuritynon-functional testingPHPAIrisk managementTesting an AI agent: where to start?https://ismail3000.fr/en/articles/tester-un-agent-ia-par-ou-commencer/https://ismail3000.fr/en/articles/tester-un-agent-ia-par-ou-commencer/An AI agent never answers exactly the same way twice. Here is how to build a first test strategy that holds up: scenario sets, evaluation criteria, guardrails and regression testing.Sat, 24 Oct 2026 07:00:00 GMTAI & QAAI agentsLLMtest strategyISTQB AI TestingStartup without QA: the first 5 quality habitshttps://ismail3000.fr/en/articles/startup-sans-qa-5-premiers-reflexes/https://ismail3000.fr/en/articles/startup-sans-qa-5-premiers-reflexes/No QA team, developers shipping fast and bugs reaching your customers? Five simple habits to put in place this week, without hiring or slowing down.Sat, 24 Oct 2026 06:00:00 GMTStartupsstartupstest strategyCI/CDMoving mobile tests from Appium to Maestro: my takeawayshttps://ismail3000.fr/en/articles/migrer-de-appium-vers-maestro/https://ismail3000.fr/en/articles/migrer-de-appium-vers-maestro/I led the migration of an Appium mobile test suite (Java, Cucumber) to Maestro. Why switch, how to organise the move, and the traps to avoid.Sat, 24 Oct 2026 05:00:00 GMTToolsMaestroAppiummobile testingautomationYAMLLLM-as-a-judge: having one AI evaluate another AI's answershttps://ismail3000.fr/en/articles/llm-as-a-judge-evaluer-un-assistant-ia/https://ismail3000.fr/en/articles/llm-as-a-judge-evaluer-un-assistant-ia/When you can no longer compare word for word, you can ask an LLM to grade the answers. Powerful, but full of traps: here is how to use it without fooling yourself.Sat, 24 Oct 2026 04:00:00 GMTAI & QALLMLLM-as-a-judgeevaluationAI agentsAutomation KPIs you can actually present at a sprint reviewhttps://ismail3000.fr/en/articles/kpis-automatisation-sprint-review/https://ismail3000.fr/en/articles/kpis-automatisation-sprint-review/The number of automated tests says nothing about quality. Here are the metrics I track and present at sprint reviews to steer test automation, and the ones I avoid.Sat, 24 Oct 2026 03:00:00 GMTMethodsKPIsautomationagilesprint review