[scorecard][aggregation] Reduce unsupported aggregate-stage reasons from failure ledger

[midagedev]

Parent Epic: #98

Why

unsupported UTF aggregate stage in pipeline remains in top unsupported reasons.

Scope

• Use latest failure ledger to identify the top N unsupported stages/operators with highest impact.
• Implement the highest-value subset for in-memory test usage.
• Keep unsupported behavior deterministic for remaining stages.

Success Criteria

• Aggregate-stage unsupported count decreases in rerun scorecard.
• Imported cases increase while keeping mismatch/error at zero.
• Compatibility docs updated with newly supported stage/operator subset.

Priority

P2 (can run in parallel with CRUD/UTF adapter work).

[midagedev]

No direct stage-support implementation landed for this issue in PR #105.

Current local unsupported reason summary still contains:

• unsupported UTF aggregate stage in pipeline: 28

This issue remains the next aggregation-focused gap after bulkWrite/transaction mismatch buckets are reduced.

[midagedev]

참고: 복잡 쿼리 인증 트랙 #237 생성됨.
join-heavy/조합 파이프라인 인증 이슈 #241은 #104(aggregation unsupported reduction)과 범위 조정이 필요합니다.
중복 구현 방지를 위해 #104는 broad unsupported 축, #241은 certification subset 축으로 정리 권장.