Mutation Testing as a Merge Gate for Agent-Written Tests

<p>A patch touches zero test files. No assertion changes, no new skips, no fixture edits. The suite goes green on the first run.</p> <p>Then someone opens the job duration graph

<p>92% code coverage. No SonarQube criticals. All green. And an AI-generated deduplication bug made it to production because not a single test had challenged the logic.</p>

<p>In Part 2 of this series, we looked at boundary mutations — the category with the highest detection rate (63.8%). The numbers were reassuring, with a catch: 36.2% of boundary

<p>A recorded expected value is a leak. An agent that can read <code>assert f(x) == y</code> can patch <code>f</code> until that line is green and leave every unlisted input

<p>I have a small Python library that parses CSV files and extracts course schedules. It has no tests. I know, I know. So last week, I decided to see if a free model could write

<p>Built a .NET invoicing library with Claude Code over about a week. Zero dependencies, VAT and GST arithmetic, JSON and CSV export, the usual. By the third phase I had 202 tests

<p>An AI coding agent implements a feature, generates tests, runs them, and opens a pull request with a green build.</p> <p>That looks like independent verification. It often is