WWaseem Anjum37m ago · 11 min readHow to Test an MCP Server: Vitest, In-Memory Transport, Streamable HTTP and the Inspector CLITesting an MCP server means checking three layers: the pure functions behind your tools, the MCP contract (tool list, schemas, results and errors), and the transport your clients actually use. You canfabwaseem.hashnode.dev
AAIDevBuilds1h ago · 16 min readYour LLM Judge Agrees With Itself. That Is Not an Eval.A judge model can give the same verdicts run after run and still be wrong. In a June 2026 preprint on judge reliability [1], Qwen 3 8B reproduced its own scores almost perfectly on MT-Bench, with a teaidevbuilds.hashnode.dev
FFightNome17h ago · 2 min readTesting AlarmKit alarms in the iOS 26 Simulator: what you can and can't captureI build FightNome, a boxing and Muay Thai training app, and this note comes from adding alarms to it. We added workout alarms in iOS 26 with AlarmKit. You set an alarm on a workout with a time and dayfightnome.hashnode.dev
SSergii Davydov10h ago · 8 min readYour Bug Reports Are Lossy CompressionQA engineers spend their days reproducing bugs that were already reproduced once. The problem isn't skill. It's that we throw away the evidence. Here's a loop every QA engineer knows by heart. You finsergiidavydov.hashnode.dev
AAshen Jayawardena10h ago · 23 min readThe Tester's Guide to RAG Evaluation, Part 2: Building the EvalsIn Part 1 you met the RAG system: retrieval finds passages in a corpus, generation writes an answer from them, and each half fails in its own way. You spiked the system to see how it really behaves, aashenasserts.hashnode.dev
FFelipe Maschio16h ago · 7 min readHow I got a thread dump from a timed-out GitHub Actions testA test hangs in CI, the job sits there for 30 minutes, and GitHub Actions cancels it. The log says the job exceeded its time limit and nothing about what the test was doing when it got stuck. Running maschiojv.hashnode.dev
Jjaberadam20011d ago · 9 min readWhy Your AI Agent Calls the Wrong Tool (and How to Test Tool Routing)TL;DR: An AI agent doesn't look tools up by name. On every turn it reads every tool description and picks the closest match for the request. That makes tool selection a relative decision, so adding on36KGDsimplyexplained.hashnode.dev
AAshen Jayawardena15h ago · 22 min readThe Tester's Guide to RAG Evaluation, Part 1: FoundationsIf you test software, you already ask the question that matters most in AI evaluation: how would I know this is wrong? Now ask it of a system that answers questions from a company's documents. The samashenasserts.hashnode.dev
BBledar Gjata16h ago · 5 min readThree refusals, two verified actions: five evidence-bound Rust cycles in ESS-MAIA useful governance experiment should make refusal inspectable and successful execution conditional. ESS-MAI POC011 asks a small question: can distinct computational duties preserve missing evidence agjatalegacy.hashnode.dev
BBenchTested19h ago · 5 min readClickUp Review 2026: Free Plan Tested on Windows (8 Steps, 7.4/10)This is a republished copy of our review. The original, with all 25 screenshots, the full pass/fail checklist and pricing, lives at benchtested.app/apps/clickup. Verdict: 7.4/10, Recommended. Best forbenchtested.hashnode.dev
