# Case 03: Exact Lean badge from a floating exclusion

Self-authored synthetic review scenario, not a published software claim or customer assertion. The source material is public and pinned to `fd4aeeb2ee4fc729c18d98444fed42fd0529eeeb`.

## Proposed claim

A Python implementation is Lean verified to establish exactly three mutually unbiased bases in dimension six.

Review the [declared model and statuses](input.json) against the source. Cases 09/10 deliberately supply synthetic statuses to test documentary completeness. Those statuses are not authenticated evidence or proven hypotheses. Other cases explicitly leave operational obligations unknown.

## Source pointers

- [lean/docs/266.md](https://github.com/openai/math/blob/fd4aeeb2ee4fc729c18d98444fed42fd0529eeeb/lean/docs/266.md), SHA-256 `3141714e25026b26a4028e27ee001e70295310bad4290b5a8f7c9fd546e8a174`.
- [preprints/The-maximum-number-of-mutually-unbiased-bases-in-dimension-six-September-24-2026/The-maximum-number-of-mutually-unbiased-bases-in-dimension-six-September-24-2026.pdf](https://github.com/openai/math/blob/fd4aeeb2ee4fc729c18d98444fed42fd0529eeeb/preprints/The-maximum-number-of-mutually-unbiased-bases-in-dimension-six-September-24-2026/The-maximum-number-of-mutually-unbiased-bases-in-dimension-six-September-24-2026.pdf), SHA-256 `b4cc74c840c902494d589b51c8e9706490f37cf2735e69a0486122131c6b7446`.

Pointers identify exact files, not a new complete-file reading, theorem proof or selected-run certificate. Determine the statement, hypotheses, formal coverage and missing model/implementation bridges from the source itself.

## Review fields

Record preparation seconds, material gaps with exact evidence locations, unsure items, false alarms after adjudication and correction seconds in the [blank result template](../review-results-template.json). Do not open the report/key during the source-only arm. No expert has completed either arm.