Why AI Agents Write Weak Tests for Their Own Code: Single-Agent TDD vs. Two-Agent Blind TDD (0% vs. 100% Mutant Kill Rate)
💡 AI Labo Benchmark Series — Experiment E06 (Verification & Mutation Testing) This post is part...
Tag archive
💡 AI Labo Benchmark Series — Experiment E06 (Verification & Mutation Testing) This post is part...

Researchers found AI-built apps handing private data to visitors who never logged in. Tests that only ask as the owner would not catch that. Here is the test that would.

How I build with LLM-TDD, shown on WhatYouSay: evals lifted a voice-note transcriber from 69.9% to 86.4% word match, and every live failure becomes a check.
En bref La répartition des tests dépend de ce que le projet peut casser. Un site sans backend se...
Two 2026 empirical studies land the same shape: tests generated from a task description alone catch roughly 25% of injected faults, while tests generated after the model has seen the implementation catch roughly 14%, and chain-of-thought and chain-of-verification prompting do not close the gap. Red-green-refactor's temporal order was never a workflow preference. It was a provenance claim about the test, and the agentic workflow that generates code and then back-fills tests inherits a measurable disadvantage the human discipline was always built to prevent.
I write most of my code with Claude Code these days. For a long time my workflow was built on...
In your first pairing session on the new team, your colleague opens an empty test file before...
A ticket shipped end to end with Claude Code: one model per step, a test seen red before the code, proof by undoing the fix, and a review that stops on a written condition.

A practical AI-driven development workflow using TDD, multi-agent code review, MCP, and Kotlin guardrails like Detekt, Kover, PMD CPD, and Konsist.
"It works on my machine" is a frustrating way to find out that an ETL job depends on a particular...

In just a few years, artificial intelligence has moved from the lab to our pockets, and then into our...

You probably remember how hard it was to explain to the business why we needed to set aside a...