Validate CSV size, headers, and cell limits before Firestore import

[MoerAI]

Summary

CSV import has no visible limits or validation for file size, row count, column count, header shape, duplicate/empty headers, or cell size before parsing and writing to Firestore.

Evidence

• src/lib/utils/csvParser.ts:9-28 parses arbitrary CSV with Papa.parse and returns headers/rows directly.
• src/app/(main)/admin/tasks/create/page.tsx:37-53 accepts the selected CSV without size/shape checks.
• src/app/(main)/admin/tasks/create/page.tsx:153-159 maps all parsed rows into Firestore row payloads.
• src/lib/firebase/firestore.ts:231-257 batch-writes rows to Firestore.

Impact

Large or malformed CSV files can cause browser memory issues, excessive Firestore writes/cost, documents exceeding Firestore limits, silent overwrites from duplicate headers, and corrupted task data from empty headers or unbounded cell values.

Minimal Fix

Add validation before creating any task/rows:

• maximum file size
• maximum row count
• maximum column count
• maximum header length
• maximum cell length
• reject empty headers
• reject duplicate headers
• reject or normalize invalid rows before write

Also consider enforcing schema/field limits in Firestore rules where possible.

Acceptance Criteria

• Oversized CSV files are rejected before parsing or before Firestore writes.
• CSVs with empty or duplicate headers are rejected with a clear error.
• Excessive rows/columns/cell lengths are rejected with a clear error.
• A failed CSV import leaves no partial task document or partial row data behind.