Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Bank Reconciliation

    1

    Month-end bank reconciliation, done by your agent in seconds and checked by you.

    $24

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Bank Reconciliation

    Bank Reconciliation

    Example session with this skill installed

    Using the Bank Reconciliation skill: reconcile this bank statement against my open invoices and supplier bills and tell me what needs review.

    • Read your context and instructions
    • Compiled the bank reconciliation

    Summary

    I ran the included sample files (bank statement, 8 open invoices, 3 supplier bills):

    Customers: 5/6 deposits auto-matched (83.3%)
      1 deposits need review, 2 invoices still unpaid
    Suppliers: 2/3 debits auto-matched (66.7%)
      1 debits need review, 1 bills still unpaid
    Wrote reconciliation.xlsx
    

    What matched, and how

    Bank lineMatched toType
    PAYMENT INV-2001 THANK YOU 1,500.00INV-2001ref_in_description
    EFT FROM GREENFIELD TRADING 2,350.50INV-2002unique_amount
    BULK EFT SETTLEMENT 1,800.00INV-2004 + INV-2005split_payment
    PAYMENT 30045 SILVERLINE 720.0130045 (1 cent off)ref_in_description

    Needs review: UNKNOWN DEPOSIT 415.15 (the closest invoice is 415.00, 15 cents apart, so it is left for you).

    Next steps

    • Check the yellow (medium-confidence) matches, then post the matched items in your accounting system.
    • Use --brand "Your Firm" to put your own name at the top of the report.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Month-end bank reconciliation, done by your agent in seconds and checked by you.

    What it does

    • Money in is matched to customer invoices; money out to supplier bills (optional).
    • Four passes: invoice number in the description; a unique amount; customer name plus amount; one payment settling 2 to 3 invoices.
    • If two or more invoices fit and nothing separates them, it leaves the payment for you instead of guessing.
    • A formatted Excel report with your own firm's name: matched items colour-coded by confidence, unmatched deposits, unpaid invoices, unmatched debits, unpaid bills, and a list of any rows it could not read.

    Built to be trusted

    • 27 checks, all passing, including one for each of: a payment exactly 1 cent off, digit-only invoice numbers, short or blank references that could cause false matches, unreadable rows, supplier bills, and Excel formula injection (a payer name like =HYPERLINK(...) is stored as plain text).
    • 1,500 payments against 1,500 invoices in 1.3 seconds on my PC.
    • Runs locally. No network calls. Nothing is uploaded.

    Honest limits

    • Tested on invented data and CSV files. PDF reading has not been tested on real bank PDFs, and scanned (image) PDFs are not supported.
    • Partial payments, credit notes and foreign currency are not handled.
    • Medium-confidence matches are suggestions: a person must review them before posting anything.

    Licence: for your own business and your clients' work; no resale.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.

    No support for scanned PDFs, partial payments, credit notes, or foreign currency. Requires Python 3.8+ with pandas and openpyxl installed locally.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 16 days ago

    Software
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Local only. Terminal/Shell runs scripts/reconcile.py. Read Files reads the bank statement, invoices and bills you point it at. Write Files writes the one Excel report you name. No network access, no environment-variable reads, no telemetry. Descriptions starting with = + - @ are written as plain text so they cannot run as formulas.

    Listed16 days ago

    What's inside

    Frequently Asked Questions

    The skill requires a standard bank statement (CSV or text-based PDF) and an export of your open invoices and bills in CSV format. It is designed to work with data from any accounting software, including QuickBooks, Xero, or custom ERPs.

    Unlike a basic LLM prompt, this skill uses a deterministic Python engine that follows strict financial logic. It handles complex tasks like split payments and penny tolerances that often cause AI models to hallucinate or make calculation errors.

    Yes, the logic runs locally on your machine via Python, meaning your sensitive financial data is processed privately and is never sent to a third-party server or used to train external models.

    You will receive the full Python source code for the matching engine, a configuration guide for setting your matching tolerances, and a sample data template to help you format your reports correctly.

    The skill identifies matches based on exact reference numbers, exact amounts, and name-based keywords. If a transaction is ambiguous, the skill flags it in a "Needs Review" Excel sheet rather than making a guess, ensuring 100% accuracy in your final ledger.