Self-hosted tools / install, operate, accept

Place a small web tool on your site: install, operate, accept

Build the existing self-hosted CSV widget, place its standalone package on an allowed static site, and verify the workflow with synthetic CSV before deciding whether to handle real data.

Build a widget See the 6 steps

Fix the boundary first

Use only the existing index.html, widget.js, and README.txt. No server, account, database, external submission, analytics, customer data, or secret is required.

Check	Accept when	Stop when
Package	Three files are together	Unknown script or secret appears
Processing	CSV is validated, corrected, and saved in the browser	It sends to a server/API
Data	Synthetic CSV reproduces the path	Customer data is used as an example

The six install → operate → accept steps

Open the widget builder and choose CSV intake.
Confirm the five mappings: submission_id, record_id, status, field, and error_message.
Save the standalone ZIP and extract index.html, widget.js, and README.txt.
Place the three files on an allowed static site and open the page in your browser.
Use the synthetic CSV; correct one failed row, explicitly select it, and revalidate.
Compare with the expected output; verify reload clears the session and no external request is needed.

Run one synthetic CSV check

sample.csv contains success, failed, and unknown states. Correct the failed row, select only that row, and revalidate. Do not guess that an unknown row is successful; ask its owner.

sample.csv → correct failed row → select one row → Revalidate → save selected CSV

The fixture uses fictional values. Do not paste real data, email addresses, tokens, or cookies.

Acceptance checklist

Three files have their specified names
Only the selected row matches the expected output
Reload or close clears the input and result
Input and results are not sent outside the browser
The owner knows how to remove the package

Save the acceptance-checklist.md before using real data.