Why a custom skill beats a one-off prompt for a weekly bank summary
Bank spending is a weekly thing. Same recurring charges, same vendors, same patterns. A personal skill knows your categories, knows that "Paybox Shiraz" is mom, and that your default fuel stop is the Sonol on the corner. A one-off prompt knows none of that and returns a generic result.
What makes this skill specifically yours
- Your categories: not a fintech app's. For you "subscriptions" includes Spotify, Netflix, AppSheet, and car insurance. For someone else it looks completely different.
- Your regular vendors: your usual grocery, your default fuel stop, your favorite cafe. The skill classifies without asking.
- Your tolerance for anomalies: an extra 50 shekels on groceries, statistical noise or warning light? Depends on the person, the budget, and the month.
Three things to include in your description
- Your CSV format: first time, open the file in Excel or Numbers and describe the columns. Banks differ in layout, so the skill needs to know yours specifically.
- Your categories list: 5-10 categories you use, each with 3-5 example vendors.
- The moment you review: Sunday morning with coffee? Thursday evening? The summary shape should match when you will actually read it.
What this skill is NOT
This is not a fintech tool, not MoneyForce, not Riseup. It does not connect to your bank, and it does not give financial advice. It reads a CSV you download, classifies into your categories, and returns a digest for human review. For full categorization automation, our catalog has options.
Drop our catalog a look first
If you want a generic spending-classification skill maintained for everyone, check the agentskills.co.il/en/categories/tax-and-finance catalog. The generator is for the parts that are uniquely yours.