fix(dashboard): enforce advanced-sequence benchmark

[grant-mccurdy]

Summary

• set the synthetic Beyond Core Math Sequence proficiency reference to 75, above all core-sequence course references
• document the expected-sequence rationale and retain explicit illustrative/non-policy language
• enforce the relationship in source publication verification
• refresh the checksum-backed dashboard manifest and SQL extract timestamp

Validation

• make check
• make privacy-check (0 failures; expected small-synthetic-group warning only)
• 4,018-record publication checksum verified
• git diff --check

Public readiness

All values remain synthetic and illustrative. The change does not claim an institutional proficiency policy or use private assessment records.

[grant-mccurdy]

Reviewed the exact four-file patch and passing publication check. The 75-point post-core reference is higher than every core-sequence reference, the verifier checks the generated dashboard rather than only source constants, the rationale remains explicitly synthetic/non-policy, and privacy validation reports zero failures.