import-statements

Parse arguments from $ARGUMENTS. Determine the data directory: use --data-dir if provided, otherwise default to ~/Documents/finance-data/.

Accept the positional argument as a file path or directory path containing bank statement files. If no path is given, ask the user to provide one before proceeding.

Format detection: For each file to process, determine the parser based on the file extension:

• Extension .mt940 or .sta → MT940 format
• Any other extension (.csv, .txt, etc.) → CSV format

If --bank is specified, pass it to the CSV parser to apply bank-specific column mappings (supported values: abn_amro, ing, bunq). If the bank cannot be detected automatically from CSV headers, prompt the user to specify --bank.

For each file, run the appropriate parser:

MT940:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/parse_mt940.py <file> [--data-dir PATH]

CSV:

uv run ${CLAUDE_PLUGIN_ROOT}/scripts/parse_csv.py <file> [--bank BANK] [--data-dir PATH]

Collect results across all files. After all files are processed, report:

Bank	File	Transactions Added	Date Range	Parse Errors

• Transactions Added: count of new rows inserted into the database
• Date Range: earliest to latest transaction date in the file
• Parse Errors: number of rows skipped due to parse failures; list the specific row numbers or error messages if any exist

Duplicate detection: If the parser reports that a transaction ID was already present in the database, warn the user explicitly:

Warning: N duplicate transaction(s) detected in <filename> — already in database, skipped.

Do not silently skip duplicates. Show the transaction IDs that were skipped so the user can verify.

If a file does not exist or cannot be read, report the error and continue with the remaining files.