defect-reportlisted

# defect-report — the anomaly, written so it can be acted on A red test says *something is wrong*. A defect report says *what is wrong, how to see it again, and against which promise*. QAIA produced the first and never the second, which is the gap this skill closes ([#75](https://github.com/QAIA-Project/QAIA/issues/75), the backlog entry that named it). ## The rule that governs everything here **Never state a cause you have not observed.** A report that asserts an unverified diagnosis is worse than no report: it sends a developer down a path chosen by a model, and it costs more than the defect. Every sentence of the output must be readable in the attached evidence — the trace, the response body, the screenshot, the diff of expected vs actual. Where the cause is unknown, the report says so, in those words. This is the same rule as everywhere else in QAIA (`references/evidence-discipline.md`), and it is the one that most often gets bent under pressure to sound useful. ## Inputs | Input | Required | Where it comes from | |---|---|---| | The failing test's output | yes | Playwright reporter, `results.json`, terminal | | The scenario it implements | yes | the `@QAIA-xxx` tag in the test title → the `.feature` | | The requirement behind it | yes | the test book's stated source (see `REQUIREMENT-SOURCE.json` if present) | | Trace, screenshot, HAR | if produced | `test-results/`, `trace.zip` | | The system under test's version | yes | commit, tag, or build identifier | If the