Ready-to-paste answers for the Developer Dashboard. Everything here matches the code as shipped — if the code changes, re-check these before resubmitting.
Upload artefact: releases/roose-flow-1.0.0.zip (npm run package).
Name
Roose Flow
Summary (132 character limit)
Queue image prompts from a text file, generate them in batches, and save the results automatically.
Description
Roose Flow turns a list of prompts into a hands-off batch run.
Load a .txt file of numbered prompts, review the queue, and start. Roose Flow
enters each prompt in turn, waits for the render to finish, saves it, and moves
on to the next one.
FEATURES
• Import numbered prompts from a .txt file — multi-line prompts stay intact
• Review the queue before starting, and remove any prompt you don't want
• Add prompts by hand as well
• Resume from any prompt number, so an interrupted run costs you nothing
• Automatic retries when a generation fails, with the error caught in seconds
rather than waiting out a timeout
• Automatic download of finished renders to a folder you choose
• Configurable random delay between generations
• Runs in Chrome's side panel, so refreshing the page won't interrupt a run
PRIVACY
No accounts, no servers, no tracking. Your prompts stay in memory on your own
machine and are discarded when you close the panel. Settings are stored locally.
Nothing is ever uploaded.
Roose Flow is an independent tool and is not affiliated with, endorsed by, or
sponsored by Google.
Category: Workflow & Planning Language: English
Single purpose
Roose Flow has one purpose: to run a user-supplied list of image prompts as a
batch on flow.google.com, entering each prompt in turn and saving the resulting
render to the user's computer.
Permission justifications
| Permission | Justification |
|---|---|
storage |
Stores the user’s own settings (download folder, delay range, auto-download toggle) locally, and holds the prompt queue in memory while the panel is open. |
downloads |
Saves each finished render to the folder the user chose. This is the extension’s core output; without it the user would have to save every image by hand. |
sidePanel |
Renders the extension’s interface — prompt queue, settings, and run controls — in Chrome’s side panel. |
debugger |
Sends a single real mouse click on the Flow page’s send button. The page ignores clicks produced by a page script, so a genuine click has to come through Chrome’s debugger protocol. Attached only to the Flow tab, only while a batch is running, and detached when it stops. No page data is read through it. Expect extra scrutiny at review for this permission. |
https://flow.google.com/* |
The only site the extension operates on. Needed to type each prompt into the page’s prompt box, submit it, and detect when the render has finished so the next prompt can start. |
Remote code: No. All code ships inside the uploaded package; nothing is fetched or evaluated at runtime.
Data usage: answer no to every data-collection category — the extension collects nothing — then certify:
Privacy policy URL: host PRIVACY.md publicly (GitHub Pages or a gist) and
paste that URL.
These cannot be generated from the code:
The 128×128 store icon is public/icons/icon-128.png, produced by npm run icons.
PRIVACY.md and copy its URLnpm run package, then upload the zipREADME.md