Evaluate a rule without turning unknown evidence into false
Declare how true, false, and unknown combine, then keep an unresolved result separate from permission to act.
Kiguri Blog
Audit rotating remainders in an equal whole-unit allotment
Track whole-unit base shares, a rotating extra, and cumulative quantities across a declared equal-allotment cycle.
Check changeover cost before batching preparation work
Keep task service and type-switching overhead separate when comparing invented preparation sequences.
Count reusable preparation stations from fixed occupancy intervals
Use fixed occupancy intervals and an explicit release rule to check how many interchangeable stations a fixture needs.
Check indivisible preparation blocks before pooling capacity
Test whether uninterrupted blocks fit individual sessions even when aggregate demand equals available capacity.
Check a preparation sequence against each job’s deadline
Compare job-specific completion and lateness across two fictional preparation orders without changing the work.
Challenge a check with deliberately flawed example outputs
Use a seeded-defect register to see which meaningful errors pass a check and whether harmless changes remain acceptable.
Check order effects before preferring one draft
Separate draft identity from first or second position when inspecting a small fictional A/B comparison.
Reduce a failure example without claiming its cause
Keep a deletion trace that separates a small reproducing example from a diagnosis or a globally smallest input.
Check a text record survives encoding and decoding
Compare exact rows, fields, and values after a known local CSV round trip, including a deliberately broken representation.
Compare file content without trusting names or dates
Use harmless local bytes to distinguish content evidence from filenames, displayed meaning, and approval.
Audit a two-table match before trusting its row count
Trace zero, one, and multiple matches per source identity instead of accepting a reassuring final total.