RFC Requirements Collector — privacy

An internal Smart4Aviation tool for collecting Requests For Change. Last updated 7 September 2026.

This tool interviews the person raising a Request For Change (RFC), helps them write it down, and files the result as a ticket in Smart4Aviation’s Jira. This page says exactly what it does with the data involved. It is not a marketing page and there is nothing here about advertising, profiling or selling data, because none of that happens.

Who runs it

Smart4Aviation (S4A). Questions, corrections and deletion requests: arkadiusz.janik@smart4aviation.aero.

Signing in

You sign in with your Atlassian account. We ask Atlassian for one thing — the read:me scope, which returns your account id, your name and your e-mail address. We do not ask for access to Jira, Confluence or any other product on your behalf.

• The access token from Atlassian is used once, immediately, to read that identity. It is not stored and never written to disk.
• S4A staff are taken through the company Google login by Atlassian itself. That step happens inside Atlassian; this tool does not see your Google password or account.
• Your e-mail address is then checked against a list of people allowed to file RFCs, kept by S4A in a Google Sheet. From that list we read your name, surname, your organisation’s ICAO code and your Jira account id.

What is stored, and where

On the server that runs this tool (a local database file and an uploads directory):

• your e-mail, name, surname, ICAO code and Jira account id, copied from the S4A list;
• everything you type into the form, and every revision of it;
• documents you attach, as files;
• your conversation with the assistant;
• an audit trail: what you wrote, what the assistant proposed, what you confirmed and when, plus the calls the tool made on your behalf.

The audit trail exists so that an RFC can always be traced back to what the requester actually said, rather than to what somebody later remembered.

Who else sees it

• Atlassian (Jira Cloud) — when you approve the draft, the RFC is created as a Jira issue in S4A’s instance. The ticket carries the eleven items of your request and, as attachments, the documents you supplied, a transcript of the conversation and the audit trail.
• Anthropic — the text of your request and your attached documents is sent to the Claude API, which is what reads, structures and reviews it. Anthropic processes it to return the answer; it is not used to train models.
• Amazon Web Services — only when a documentation check is explicitly requested, a short question about S4A product documentation is sent to an S4A service running in AWS.
• Google — the sheet with the list of authorised users is read from Google Sheets using an S4A service account. Nothing you type is written there.

Nothing is shared with anyone else, and there are no third-party analytics, trackers or advertising scripts on any page of this tool.

Cookies and browser storage

• one session cookie (rfc_session), HttpOnly, holding your signed-in e-mail and an expiry — it is what keeps you logged in;
• a short-lived cookie during sign-in that protects the login step against being triggered from another site;
• your current draft id in the browser’s localStorage, so a refresh does not lose your work. “Sign out” clears it.

How long it is kept

Drafts, uploads and the audit trail stay on the S4A server for as long as the RFC record is useful to S4A engineering; Jira tickets and their attachments follow S4A’s own retention rules for Jira. If you want a draft or an upload of yours removed, write to the address above and say which one.

Your choices

Using this tool is voluntary — an RFC can always be raised by writing to your S4A contact instead. If you do not want your request processed by the Claude API, do not use this tool and say so to your S4A contact; they will take the request down by hand.

This instance runs on 127.0.0.1 for development. When the tool is deployed for real use, this page moves with it and the address changes; the substance does not.