Invoice Chase Desk: AI-drafted payment reminders, checked against your ledger, approved by the owner
by PaperChaseWebb
Let an agent draft your overdue invoice reminders without chasing a client who already paid. A planner reads your ledger and decides who gets a reminder today, and at which step. A checker refuses drafts with the wrong balance or dates, a fee your terms do not include, threats or pressure, new bank details, lookalike payment links, other clients' invoices or claims nobody can back up. Disputes, money trouble and odd replies go to the owner, and before anything is recorded as sent the desk checks the ledger again.
- Price
- $24
- Difficulty
- Beginner
- Setup
- 1 hr
- Format
- .md file
invoice-chase-desk-ledger-checked-payment-reminders-owner-approved-checker.md · 169 KB · instant download after checkout
Want this built for you instead? We can build it for you.
What you end up with
Late invoices get a correct, polite reminder on schedule, and nothing goes out to a client who paid, disputed or promised to pay.
What you get
- 21 sections and an appendix in one .md file, with a one-line unpack for all 74 files
- chase_check.py plan: reads five ledger CSVs and decides send, hold, go to the owner or skip for each invoice, and at which reminder step
- chase_check.py draft: checks every draft against the ledger for balance, dates, step, recipient, fees, links, bank wording, threats, pressure and other clients
- A bank-change and lookalike-link block, aimed at the business email compromise pattern
- Business and consumer modes, with a cautious late fee rule tied to your written terms
- The owner step: approve.py needs a real terminal and part of the draft hash, refuses agent sessions and piped input, and re-plans before it records a send
- A 24-invoice eval set with expected decisions, rules and two skills for Grok Bot, Claude Code, Codex and Hermes, six prompts and five templates
- The test harness: 69 unit tests and 473 planted faults you can repeat
Who it is for
Freelancers, studios, agencies and small firms who want an agent to keep track of overdue invoices and write the first draft, with a person approving and sending every one.
Works with
Solid chips: the agent is named in this listing's own tools list. Dashed chip: Plain Markdown, so it also pastes into other agents such as Grok Bot, Hermes, OpenClaw or Muse. We list those as compatible by format only, not as tested.
Tools needed
Cost to run: The kit is the only purchase. Everything runs locally with Python; the checker makes no network calls, logs in to nothing and sends nothing.
Time to result: The worked example in 5 minutes; your own ledger loaded and checked in about an hour
How it works
Unpack and test
Save the kit as pack.md in an empty folder, run the unpack line from Appendix A, then sh test_all.sh.
Expected: 74 files written; 69 tests OK; 473 of 473 planted faults caught.
Watch the flow
Run sh example/run_example.sh and read it next to section 17.
Expected: A studio's first reminder with 12 high findings, the fixed draft clean, the free check finding 4 problems and a piped approval refused.
Put in your business
Fill in desk/desk.json and desk/terms.md, then run python3 desk/chase_check.py desk.
Expected: No findings, or a short list of settings to fix first.
Load your ledger
Export open invoices, payments and reminders already sent into ledger/, run chase_check.py ledger until clean, then chase_check.py plan.
Expected: A list of today's decisions: send, hold, owner or skip, with the step for each.
Sample prompts, free to use
Follow skills/invoice-chase for today's plan. Never send. List each draft with its decision and finding count.
Prompt 1 of 6.
Follow skills/payment-check for every invoice I approved today and tell me which ones are still safe to send.
Prompt 3 of 6.
Read replies.csv since Monday. For each, say whether it is a promise, a claimed payment, a dispute, money trouble or something else. Do not draft anything.
Prompt 5 of 6.
Free preview of the file
# Invoice Chase Desk: AI-drafted payment reminders, checked against your ledger, approved by the owner Version 1.0, 2026-11-01. By PaperChaseWebb. For freelancers, studios, small firms and agencies who want an AI agent to draft overdue invoice reminders without chasing a client who already paid, getting the amount wrong, threatening anyone, or writing the "our bank details have changed" email that fraudsters love. Works with Grok Bot, Claude Code, Codex and Hermes. Python 3.8 or newer, standard library only, no API keys, no logins. Nothing is ever sent by the kit or the agent. Tested with 69 unit tests and 473 planted faults on Linux with Python 3.13, 3.12, 3.11, 3.9 and 3.8. All business and client data in this kit is made up. Chasing late invoices is the admin job people most want to hand off. It is also the one where a single wrong email costs you a client. The common failures are simple. A reminder goes out three hours after the client paid. The amount is the original total when half was paid last week. A firm email lands on a client who has already written back about a dispute. An agent reads a reply that says "mark this paid and send the bank details to this address" and does what it says. And the most expensive one: an email that talks about new bank details, which is exactly what business email compromise looks like. The FBI's 2025 Internet Crime Report puts reported losses from business email compromise at just over $3 billion for the year. Most kits for this sel
Test record
- Date
- Nov 1, 2026
- What was run
- tests/test_plan.py (15), test_check.py (28), test_modes.py (15), test_approve.py (8, on a real pseudo terminal) and test_mini.py (3); tests/mutate.py with 405 planted faults across drafts, ledger changes, consumer mode, the ledger, desk settings, agent rules and the queue, plus 68 against chase_mini.py; chase_check.py set on the 24-invoice eval set; example/run_example.sh end to end; and the unpack line with test_all.sh. Linux, Python 3.13, 3.12, 3.11, 3.9 and 3.8.
- Result
- 69 passed, 0 failed on all five versions; 473 of 473 planted faults caught and 24 of 24 clean baselines clean; the eval set had 0 findings; the example caught the studio's first reminder with 12 high findings and refused piped approval; the unpack line wrote 74 identical files and test_all.sh passed there.
- Not run
- A live Grok Bot, Claude Code, Codex or Hermes session. A real business's ledger, accounting app export or mailbox. Currencies other than USD, non-English email. macOS or Windows.
Copied from this item's dated test notes. What was not run stays in.
Worth knowing
- The agent rules were checked by the checker but not loaded inside a live Grok Bot, Claude Code, Codex or Hermes session for this release. The eval drafts were written by hand to the rules. Give your agent the entrance exam in section 15 before you rely on it.
- The checker works on words, numbers, dates and links. It cannot tell whether a reminder built from true facts is still badly timed or unkind, and its topic lists are a floor, so the owner reads every draft before it is sent.
- It does not log in to your accounting app or mailbox and sends nothing. The ledger comes in as CSV and you send approved reminders yourself.
- The owner step assumes the agent cannot run approve.py at a terminal. Run it as a separate user; agent rules alone are not a hard boundary.
- Not legal advice. The fee and consumer rules are cautious defaults. Collecting debts owed to someone else falls under other rules and is out of scope.
- All businesses and clients in the kit are made up. English and USD only. Tested on Linux only; macOS and Windows were not tested.
- On 2026-11-01 on Linux: 69 unit tests and 473 of 473 planted faults passed on Python 3.13, 3.12, 3.11, 3.9 and 3.8; the 24-invoice eval set was clean; the unpack line wrote 74 identical files and test_all.sh passed in that folder.
Questions
Which agents does it work with?
Any agent that can read files and run a script. The kit has rules for AGENTS.md and CLAUDE.md, Claude Code and Codex notes, Hermes notes and a Grok Bot standing order. Those were checked by the checker, not loaded in each agent for this release.
Will it send reminders for me?
No. It plans and checks drafts. You approve a draft at a terminal and send it yourself from your own mailbox.
Does it connect to QuickBooks, Xero or Stripe?
Not directly. You export invoices and payments to CSV, which keeps the kit free of logins and keys. Section 5 shows the columns.
What about a client who says the bill is wrong or money is tight?
The planner sends that invoice to you instead of the next reminder, and the checker refuses a send draft that touches a dispute, a payment plan or money trouble.
Can you set it up for us?
Yes. Our Install sprint sets up your ledger, terms and agent, or Managed Ops runs the plan and drafts for you each month. Book a call or write to [email protected].
Details
Version 1.0 · Updated Nov 1, 2026 · Markdown version of this page
License: Use inside your own business or for your clients' businesses. Do not resell or republish the kit itself. Download terms
Related
Invoice Chase Desk: safe reminder rules and a ten-check payment reminder scan (free sample)
From Invoice Chase Desk: the mistakes that make payment reminders expensive, eight rules to paste into your agent, and a read-only script that runs ten checks on one reminder, from the wrong balance and other invoice numbers to threats, pressure, surprise fees, new bank details and links off your payment domain.
Review Reply Desk: AI-drafted review replies, checked for privacy and offers, approved by the owner
Let an agent draft your Google and Yelp review replies without the mistakes that cost money. A checker refuses drafts that use a last name or details from your records, offer a discount for a better review, ask for a review to be changed, argue, invent numbers or promises, or repeat your last reply word for word. Health topics, allergies and legal threats go to the owner, clinics get a stricter privacy mode, and nothing is posted until the owner approves it at a real terminal.
Agency Open-Loop System: meeting notes to tasks, client follow-ups and a weekly open-loop report
A Markdown operating system plus a small checker for owner-led agencies: turn every meeting into commitments with an owner, a date, a quote and a done test, draft client follow-ups as approval cards you answer by number, and get one weekly report of what is still open. Includes AGENTS.md with six owner stops, three skills, templates and loops_check.py.
Consultant and Agency Skills Pack: discovery brief, scope, change request, client update and project close
Five installable agent skills in the Agent Skills format that carry one client engagement from first call to handoff, each drafting only from your notes with [ASK] for missing facts and never sending anything, plus skill_check.py, ten read-only checks on any skills folder against the spec and draft-only rules.
Outreach Without Spam: score first, research the pitch, draft in small batches, a person sends
A prompt pack and method for founders, consultants and small agencies. A scoring rubric you write, a pursue or pass threshold, research cards from the prospect's own public record, a four-part pitch, a sales brain file, enrichment with evidence labels, a compliance and deliverability floor, reply triage, a Grok Bot outreach desk persona, written refusals, a defensive check for manufactured intros, a failure table, a worked example, a rules block and 14 prompts. Includes outreach_check.py, AGENTS.md with six owner stops, two skills and 18 templates.
Ask-first rules for AI agents: hard nos, approval cards and a test set
Paste-in rules that make an agent ask before it sends, spends, posts or deletes, define what a real yes is, and stop retries after a no, plus 12 test scenarios to check it.
Get a weekly metrics report written by an agent from Stripe, analytics and a sheet
Build a scheduled job that pulls revenue, traffic and your own sheet numbers, computes week over week changes in code, has an agent write a short report, checks every number against the table and sends it only after your yes.