Add latest IBM OA questions

[tomtoto757]

Summary

• add the verified IBM Plus Mult Array, Drawing Edge, Increasing-Value Triplets Under a Threshold, Count Invalid Log Groups, Count Paginated Medical Records in a Range, Expiring Authentication Token Manager, Maximize Consecutive XOR, and Count Digit Holes Coding rows
• refresh Get Maximum Amount to its latest verified sighting date
• refresh the Aug 27 freshness boundary and update the README Coding count to 1,785
• preserve the canonical five-column rows, exact FastPrep routes, and newest-first ordering

This Stage-4 sync remains intentionally README-only and draft. The linked generated format pages are stale and must be regenerated before this PR can be ready or merged.

Verification

• focused README validator: 1,994 rows and 1,994 unique routes; zero malformed rows, pipe-contract violations, mismatched practice URLs, duplicate routes, or date-order violations
• README format counts: Coding 1,785; SQL 2; System design 158; Low-level design 29; AI coding 20
• Maximize Consecutive XOR and Count Digit Holes each appear exactly once with IBM, Coding, their exact canonical routes, and the latest valid sighting date, Aug 09, 2026
• FastPrep public APIs: both target projections match ID, title, IBM, OA, HackerRank, FULLTIME + NEW GRAD, and lastSeen [2026-08-09, 2026-02-14]; both exact FastPrep routes return HTTP 200
• python3 scripts/refresh-freshness.py — byte-idempotent on two consecutive runs (0 marker changes, 0 moves); README 455,993 bytes
• PYTHONDONTWRITEBYTECODE=1 python3 -m unittest -v scripts/test_sync_practice_formats.py — 10/10 passed
• git diff --check origin/master...dbe2c002c8c21f0a9233e0e06e393d9e682d10e6 — passed

Expected runtime impact: none; this changes README catalog data only.

Independent regression review

Independent reviewer: tech_oa_final_review (did not implement the README changes).

• reviewed head: dbe2c002c8c21f0a9233e0e06e393d9e682d10e6
• reviewed base / merge-base: 3c98ad57c68670b3cfb001a50c7818d470548683
• reviewed the complete README-only diff, including every prior IBM commit retained on the branch
• result: no new actionable defect
• structural validation: 1,994 rows and 1,994 unique routes; no malformed rows, mismatched practice URLs, duplicate routes, or date inversions
• format counts: Coding 1,785; SQL 2; System design 158; Low-level design 29; AI coding 20
• exact Qvick rows and all nine affected IBM routes, public APIs, and public pages passed
• PYTHONDONTWRITEBYTECODE=1 python3 -m unittest -v scripts/test_sync_practice_formats.py — 10/10 passed
• freshness was byte-idempotent and git diff --check passed

Known draft blocker and exit criteria

python3 scripts/sync-practice-formats.py --check exits 1 because the repository-owned generated pages are stale: formats/coding.md, formats/system-design.md, formats/low-level-design.md, and formats/ai-coding.md. The Coding-page gap was 48 rows at the parent head and is 50 rows at the final head because this commit adds the two Qvick rows to README without regenerating that page.

Keep this PR draft and do not merge until all generated pages are regenerated from one bounded public-catalog snapshot and python3 scripts/sync-practice-formats.py --check passes against the final SHA.

Remaining risk: FastPrep's managed public catalog is temporal, so readiness should use one bounded catalog snapshot and rerun generation/check immediately before merge.