Demonstration 1.0.0 · Service 0.14.5
Recorded service output
Pick a routine and a run. You will see the exact request posted to the service, the verdict it returned, the numbers behind that verdict, and the complete response body. The problem runs are the point: a routine that cannot reject anything is not a control.
Every verdict and every number here was produced by the packaged AssuranceView QC
service, version 0.14.5, and recorded on
6 October 2026. This page is a viewer over that recording. It contains no
evaluation logic and computes nothing in your browser, so what you read is what the
service under evaluation actually returned — not a second implementation that might
agree with it today and drift tomorrow.
The inputs are synthetic fixtures with known ground truth. No
patient specimen, no deposited study and no laboratory's data is represented on this
page. To reproduce any run, post the request body shown to the endpoint shown on
your own instance, or run python record_runs.py.
Evidence record
Mutation score is the gate. Faults are injected into the routine code and the
suite must catch them; a suite that cannot is not evidence, whatever its coverage. Coverage is
reported because it measures which lines executed, not which behaviour was asserted, and gating
on it rewards tests that assert nothing. Both figures come from python measure_quality.py
--mutation and the page is refused a build if they are stale, if any test fails, or if any
gate is open.
Boundaries
These are properties of the build, not promises in a document. The routes are absent from the deployed product and the test suite asserts their absence.