Topics

Testing where the coding agent is the operator and the human supervises.

Why suites go unreliable, how to measure it, and what actually fixes it.

What healing actually does, the mechanisms behind it, and where it fails.

Model Context Protocol, agent skills, and the plumbing that lets agents use tools.

The framework, its agents, and the patterns that keep a suite maintainable.

Claude Code, Cursor, Codex, and how teams actually work with them.

The fundamentals: what to test, how much, and what it costs.

Running tests in a pipeline, and what a red build should do next.

Verifying chatbots, RAG and generated output when the answer changes every run.

Head-to-head comparisons and alternatives, compared on mechanism.

Security review, data handling, and testing in regulated contexts.

Test types, frameworks and the vocabulary the rest of the blog assumes.

Treating the spec as the source of truth, and what that means for tests.

Writing tests without writing code, and where that approach holds up.

Where AI actually helps in a test strategy, and where it is being oversold.

Autonomous QA loops: planning, generating, executing and evaluating tests.