Summary
A Claude Skill for accountants is a folder holding a written procedure, and often a script, that Claude loads when a task matches. The firm's method for a bank reconciliation, month-end close or debtor review gets written down once and runs the same way every time. Skills work in Claude, Claude Code, Codex and Cursor.
- What it is: Anthropic defines a skill as a directory containing a SKILL.md file of instructions plus optional scripts and resources; no model training, no data upload, no new platform.
- The one rule: The model must never do the arithmetic; every total, aging bucket and fee runs through a script, because a language model handles twenty numbers well and a hundred badly.
- The ceiling: Skills cannot watch a ledger, mostly cannot send, and handle one organization at a time, so chasing on schedule still needs accounts receivable automation.
The accounts receivable skill described in this article, in the open. Aged receivables, DSO, late fee schedules, customer statements, a ranked call sheet and verified chase briefs. Every figure is computed by a bundled script, never by the model.
Get the skill on GitHubWorks in Claude, Claude Code, Codex and Cursor. Setup for each platform is in the repository.
Interest went from nothing to serious in under a year. Our own keyword data shows searches for "claude skills" were negligible in September 2025 and now run in the tens of thousands a month, with accounting-specific variants growing behind them. Skills are available on every Claude plan, including Free, and in Claude Code.
For the wider picture of where AI for accountants helps in a practice and where it does not, start there and come back here for the skills detail.
Free: Payment reminder email templatesCopy the reminder emails for before, on and after the due date.What is a Claude Skill?
A Claude Skill is a folder. Anthropic's own definition is "a directory that contains a SKILL.md file": instructions in plain English, plus optional scripts and reference notes.
When you ask a question the skill covers, Claude loads those instructions and follows them. Anthropic calls this progressive disclosure: Claude scans the available skills, and when one matches it loads only the files that task needs. There is no model training, no data upload, no new platform.
The practical effect is that a procedure stops living in one person's head. Your best preparer writes down how your firm does a debtor review, and every other person on the team runs that exact review. New staff inherit the method on day one instead of month six.
What is the difference between a skill, an MCP server and an agent?
An MCP server is the connection to your data, a skill is the written procedure, and an agent is the assistant running that procedure against the data. The three get used interchangeably, and the distinction matters when you are working out why something does not work.
| Thing | What it is | Kitchen analogy |
|---|---|---|
| MCP server | The connection to your data. Xero and Intuit both publish one. | The ingredients and the equipment |
| Skill | The written procedure the AI follows. | The recipe |
| Agent | The assistant running the procedure against the data. | The cook |
You need all three. An MCP connection with no skill gives you an assistant that can reach your ledger but has no method, so it improvises differently every time you ask. A skill with no connection gives you a method with no data, which still works fine on a CSV export.
What do accountants use Claude Skills for?
Accountants use Claude Skills for the recurring, documented jobs that eat a morning and produce a document somebody reviews: debtor reviews, bank reconciliations, document extraction, month-end close, subledger reconciliation, chase emails and late fee schedules. Practices that want chasing to run without a prompt can use accounts receivable software for accountants and bookkeepers.
- Debtor review. Aged receivables by customer, the movement since last week, and the accounts that got worse.
- Bank reconciliation. Matching a messy statement export against invoices and flagging the variances.
- Document extraction. Reading a folder of supplier invoices and receipts into a clean file for import.
- Month-end close. The checklist, the journals, the accruals, and the review notes that go with them.
- Subledger reconciliation. Tying the AR ledger back to the control account and listing the breaks.
- Chase emails. Drafting the follow-up for every overdue invoice, in the firm's tone, at the right escalation level; credit control that uses AI does the same on a schedule, with approval before it sends (what AI handles and what a person approves).
- Late fee schedules. Working out what the contract allows on each overdue invoice, with the calculation shown.
Why must the model never do the arithmetic?
The model must never do the arithmetic because language models are poor at adding up a real ledger, and the failure is quiet. That one rule is the difference between a skill you can rely on and one that will embarrass you.
Language models are poor at arithmetic over a real ledger. They handle twenty numbers well and a hundred badly, and the failure is quiet: the total comes back looking entirely reasonable and wrong by a few hundred pounds. You find out in front of the client.
The fix is not a better prompt. It is to move every calculation into a script the skill runs. The model reads the request, runs the script, and explains the result. The arithmetic happens in code, so it is right, and it is the same every time you ask.
Three things follow from that rule, and they are what you should look for in any skill before you trust it with a client file.
What a trustworthy accounting skill looks like
The numbers come from a script, not from the model. If the skill's instructions say "calculate the total", the model is doing it. If they say "run this script", the code is doing it.
Every output shows its workings. The source file, the as-at date, the row count, and a control total the reader can check. A figure with no provenance cannot be reviewed.
Exceptions get surfaced, not swallowed. Three invoices with no email address and one with no due date are part of the answer, not a detail to quietly drop.
That last point is the one people miss. AI mistakes are hard to spot because they are not the mistakes a human would make, and a clean-looking report with two rows silently excluded is worse than no report. The skill should tell you what it could not process.
What can Claude Skills not do for accountants?
Claude Skills cannot watch a ledger for an overdue invoice, mostly cannot send anything, and handle one organization at a time. Being straight about that ceiling saves a lot of wasted setup.
They cannot watch your ledger. Nothing fires when an invoice goes overdue. Neither the Xero API nor the QuickBooks API can trigger on it. Every run starts because a person asked.
They mostly cannot send. There is no email tool in the Xero connection at all, so chase emails and statements come out as drafts for a human to send. QuickBooks can send invoices and overdue reminders through its connector, currently for United States organizations only; the detail is in our guide to Claude Skills for QuickBooks.
They handle one organization at a time. Every connector holds a single company. A bookkeeper with twenty clients reconnects twenty times, which is why the CSV export route is the one that actually scales across a book.
None of this makes skills less useful. It defines what they are: analysis and drafting on demand, not enforcement.
Is there a free accounting skill for Claude?
Yes. Paidnice publishes its accounts receivable skill as open source under the MIT license on GitHub, because the fastest way to show what a well-built accounting skill looks like is to hand over the code.
It gives you aged receivables, DSO and payment behavior, late fee schedules, customer statements, a ranked call sheet, and verified fact sheets for chase emails. It reads Xero, QuickBooks or a CSV export. Every figure is computed by a bundled Python script with no dependencies to install, and the calculations are covered by 59 tests against a fixture ledger with hand-computed answers, so you can check the arithmetic yourself in about two seconds.
It is on GitHub at accounts-receivable-skills. Setup instructions for each platform are in the repository, and there are dedicated guides for connecting Claude to Xero and connecting Claude to QuickBooks Online.
Where do skills stop and automation start?
A skill stops at analysis and drafting on demand; automation starts where the same policy has to apply to every invoice on the day it lapses. A skill is a very good analyst who only works when you are in the room. Ask it who to chase and it will tell you, with the workings. Ask it to draft eleven chase emails and it will, in your tone, with the right escalation on each.
What it will not do is notice on Saturday that an invoice has gone twenty days past terms, apply the fee your contract allows, send the reminder, and escalate when that reminder is ignored. That is not a limitation of the skill. It is a limitation of the connection underneath it, and it applies to every AI accounting integration available today.
💡 Paidnice insight
We built the skill and we still run the chasing through automation, because the two jobs are genuinely different. Judgment on demand is what an AI assistant is good at. Doing the same thing to every invoice, on time, without being asked, is what software is good at. Customers using Paidnice see half as many late invoices in the first 30 days, and none of that came from someone remembering to run a report.
Use the skill for the thinking. Use Paidnice for the doing. Paidnice is accounts receivable AI for Xero and QuickBooks, with tailored AI for your collections that reads every reply. Its accounts receivable automation runs the rest: reminders that go out on schedule, late fees that apply themselves under your rules, statements that arrive without a person triggering them, and escalation that happens whether or not anyone opened the ledger that morning. For the apps a practice runs alongside the skills, see credit control software for accountants.
Paidnice also runs its own MCP server, so you can add it to Claude or ChatGPT next to the ledger connection. The assistant can then read your customers' replies, check an invoice in Paidnice and update a customer's contact people.
Common questions
Do I need to be technical to use a Claude Skill?
To use one, no. You install a folder and ask questions in plain English. To write one, you need to understand the finance well enough to know when the output is wrong, which is a different skill from writing code.
Does my client data get used for training?
Check the terms of the specific plan you are on, because they differ between consumer and business tiers. If the question matters for your engagement letters, the CSV route keeps the scope tight: you send one aged receivables export rather than granting access to the whole ledger.
What is the difference between a skill and a custom GPT or a saved prompt?
A saved prompt is text. A skill is a folder that can carry scripts, reference files and templates, so the deterministic parts run as code rather than as instructions the model may interpret differently on Tuesday.
Can a skill post journals or send invoices?
Technically some connections allow writes, but the working convention across every credible finance skill is draft and approve. Nothing irreversible happens without a person saying yes, and that is the right default when the subject is money.
Will this replace my accounts receivable software?
No, and they solve different halves of the problem. The skill answers questions and drafts documents when you ask. The software applies your policy to every invoice without being asked.
Paidnice features in this article
About the editorial team
Denym started Paidnice in 2022 after customers paying late left his own business short on payroll. He has led it to $1 billion collected on time across more than 1 million invoices, and writes about credit control, late fees and getting paid on Xero and QuickBooks.
Ashley is a co-founder and lead developer of Paidnice and has built accounting software for more than 10 years. He builds the ledger connections that put reminders, late fees and payments into Xero, QuickBooks and ERPs such as NetSuite and Business Central. Before Paidnice he co-founded A2X, ecommerce accounting software for Xero and QuickBooks.
App of the Year