All builds
When a rule can be read two ways, which example settles it?
Turns two readings of one rule into the boundary case that tells them apart, and a test.
Design choice
Compute where the readings disagree, ask about that case, and keep the answer as a test.
In practice
Ambiguity hides at boundaries. Ask about the case where the readings differ, not the ordinary one.
What I built
- A disagreement matrix across candidate rules
- Deletion-minimal conflicts: the fewest examples that separate two readings
- The chosen reading recorded and exported as a machine-readable test
- A Python CLI and a local browser workspace
Where it stops
Human usefulness, productivity and live-model evaluation have not been run. The worked example is authored. It follows ClarifyGPT as a predecessor and is not a replication.