» Tag
testing
87 postsYour LLM Judge Needs a Regression Test Suite
rubric-bench v0.1 is an open-source tool that regression-tests LLM judges against golden sets, catching grading drift in CI before it reaches production.
Parlel Emulates 250+ Services Locally With Real Wire Protocols
Parlel runs 250+ services like Postgres, Redis, Stripe and S3 locally using real wire protocols, offering a free, fast alternative for testing and AI coding agents.
A Run ID Pattern for Stable React Auth Email Tests
Flaky React auth email tests usually stem from missing correlation, not timing. A typed run id pattern ties UI state, API, and inbox checks together for reliable assertions.
CommitBrief — AI code reviews, right in your terminal
A provider-agnostic, local-first CLI that reviews your staged changes, a historic range, or a whole GitHub pull request. Zero telemetry, no server. Free and open source.
commitbrief.comSpin Up Ephemeral Test Inboxes for Email Integration Tests
Discover the advantages of using ephemeral test inboxes for email integration testing.
How to Stop Duplicate React Mention Emails
A look at why React mention emails get sent twice and how a shared eventId contract between UI, API and worker keeps notifications idempotent.
Inbox Contracts: A Testing Pattern for Scheduled Email Jobs
A practical pattern for scheduled jobs: isolate each run with its own mailbox, run id, and evidence-based assertions to catch silent email bugs early.
Why LLM-as-judge scores flip between identical runs
An LLM judge gate flipped between 0.79 and 0.82 on identical inputs. Sampling temperature, model drift and vague rubrics cause the jitter; k-sampling and a noise band fix it.
Testing a Modular Monolith in C#: Unit and Integration Test Strategies
Explore effective unit and integration testing strategies for C# modular monoliths.
LLM Red Team Lab: An Interactive Educational Tool
The LLM Red Team Lab offers an interactive educational tool for authorized testing of local LLMs.
Sans-I/O: The Costs of Decoupling a Protocol from Its I/O Model
The Sans-I/O approach explores the implications of decoupling a protocol from its I/O model, highlighting benefits in testability and memory management alongside potential drawbacks.