- Home
- Skills
- Technical Documentation
- Dataflow Handover — where your client's data actually goes
More screenshots
Works with the AI Agents you already use
Dataflow Handover — where your client's data actually goes
A business-level data flow from your code: where data comes in, what happens to it, where it's stored, who sees it and where it leaves, every arrow traced to code and every gap settled with you before writing.
$5
Dataflow Handover — where your client's data actually goes
Example session with this skill installed
Make a data flow document for my client from this project: where does customer data go?
- Read your context and instructions
- Compiled the dataflow handover
Bean & Kettle back office: Data flow
Version 2026-11-09 · Source project code and database schema, 2026-11-09 · For the shop owner
Fictional example. Every solid arrow is traced to the included code; dashed arrows are steps handled outside this project, as stated by the developer.
Overview
flowchart LR
Customer([Customer])
Owner([Owner])
Staff([Staff])
Store{{Online store platform}}
Pay{{Payment provider}}
subgraph System["Bean & Kettle back office (this project)"]
BO[Back-office page]
DB[(Shop database)]
Report[Revenue calculation]
end
Customer -. "name, email, order" .-> Store
Store -. "payment" .-> Pay
Store -. "writes orders" .-> DB
Owner --> BO
Staff --> BO
Owner --> Report
DB --> Report
In one sentence: orders and payments are handled by the online store platform; this project holds the shop database and the back office, which lets staff view orders and the owner refund orders and check revenue.
Process 1: An order comes in
| # | What happens | Data | Stored in | Who |
|---|---|---|---|---|
| 1 | Customer checks out in the online store | 🔒 name, email; products, quantities | online store platform | customer |
| 2 | Payment is taken | 🔒 payment details | payment provider (never in the shop database) | payment provider |
| 3 | Store writes the order into the shop database | 🔒 name, email; order, lines | customers, orders, order_items | online store platform |
Watch out: order confirmation emails are not sent in this version; they are part of phase 2.
Process 2: Owner refunds an order
| # | What happens | Data | Stored in | Who |
|---|---|---|---|---|
| 1 | Owner opens an order and clicks Refund order | order number, total | — | owner only |
| 2 | Owner confirms; the page shows status refunded and lowers that month's revenue | order status | the open page only | owner |
| 3 | Owner returns the money in the payment provider's dashboard | amount | payment provider | owner |
Watch out (important): the refunded status is
not saved to the shop database in this version: it disappears when the page is reloaded. Saving refunds is scheduled before go-live. The back office does not contact the payment provider; the money is returned by hand in step 3.
Process 3: Owner checks monthly revenue
| # | What happens | Data | Stored in | Who |
|---|---|---|---|---|
| 1 | Revenue calculation reads paid and shipped orders, leaving out test accounts | orders, customers' test flag | read only | owner only (staff refused) |
| 2 | Owner opens Revenue in the back office | month, orders, revenue | figures built into the page | owner only |
Watch out: the revenue page shows figures written into the page itself; connecting it to the revenue calculation is scheduled before go-live.
Process 4: Signing in
Watch out (security): in this version the account list is readable by anyone who opens the page and any password is accepted. It is replaced by server sign-in before go-live.
Where personal data lives and travels
| Data | Enters at | Stored in | Leaves to | Who can see it |
|---|---|---|---|---|
| 🔒 Customer name | online store platform | customers.name | — | owner, staff |
| 🔒 Customer email | online store platform | customers.email | — | not shown in the back office |
| 🔒 Payment details | online store platform | payment provider only | payment provider | — |
| Staff/owner email | sign-in page | inside the page code | — | anyone who opens the page (until server sign-in) |
This project's own code sends nothing outside the system: no web requests, email sending or payment calls.
Recommended actions
| Current state | Action | Who | When |
|---|---|---|---|
| Refunds are shown on the page but not saved | Save the refunded status to the shop database | developer | before go-live |
| Revenue page shows figures built into the page | Connect the page to the revenue calculation | developer | before go-live |
| This version's sign-in accepts any password | Replace with server sign-in and stored, hashed passwords | developer | before go-live |
| Money is returned by hand in the payment provider | Record the provider's refund reference on the order | owner | from go-live |
| No order confirmation emails | Add confirmation emails | developer | phase 2 |
| Revenue months follow UTC time | Group revenue by the shop's local time | developer | phase 2 |
Delivered as dataflow.md plus a print-ready HTML (diagrams drawn); the developer appendix traces every step to file and function.
Connects securely to your tools. The creator never sees your data.
About this skill
Architecture diagrams are for developers. Your client, their auditor and their insurer want to know something simpler: where does a customer's data go?
- One flow per business process (place an order, refund, monthly report, sign-in) in plain words, with a Mermaid diagram.
- Every arrow traced to code: file and function in the developer appendix. Steps the code doesn't show are asked of you first, then drawn as you state them (e.g. "handled by the online store platform").
- Personal data marked on every flow that carries it, plus a table: enters at · stored in · leaves to · who can see it.
- External parties explicit: email services, payment providers, AI APIs, analytics, exports.
- Recommended actions: current state · action · who · when, instead of open questions.
In the included example it traces the refund flow to the exact lines: the refund is shown but not saved, and this version's sign-in accepts any password. Each becomes a current-state note with a dated action, consistent with the test report and the manual.
Every document also comes as a styled single-file HTML, print-ready for PDF. Writes in your language (English, Chinese, Japanese and more).
Works with Claude Code, Codex CLI, Gemini CLI, Cursor and other agents that can read files and run shell commands.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Technical Documentation.
Data flows that live outside the repository (another service, a no-code tool) are drawn from your answers, as steps outside the system.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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 8 days ago
- Software
Python
Permissions
Allowed Hosts
Reads your project code and schema, runs the included render script, and writes the document into your project folder. It never opens real secret files and makes no network calls.