Publish approval
By default, anybody holding the templates capability can publish. That suits most organisations and it is how Sigil has always worked.
Publish approval is a switch that changes it. With approval on, editors still do all the same work, but putting a body in front of users needs an admin.
It is off by default, and turning it on is a deliberate act. Switching it on in an organisation that has drafts in flight would strand that work behind a review nobody asked for, so Sigil will not do it for you.
Turning it on
Section titled “Turning it on”The switch lives in settings and applies to the whole organisation. It needs the settings capability. Of the built-in roles only Admin holds it, though an Admin can also give it to a custom role. See roles and capabilities.
- Open Settings in the portal sidebar.
- In the Publish approval card, switch on Require an admin to approve every signature change.
It saves as soon as you change it, and a message confirms the settings are saved. Switch it off the same way.
What changes when it is on
Section titled “What changes when it is on”| Action | Without approval | With approval |
|---|---|---|
| Edit and save a draft | Anyone with templates | Anyone with templates |
| Preview, download, test email | Anyone with templates | Anyone with templates |
| Submit a draft for review | Not offered in the portal | Anyone with templates who cannot publish |
| Publish | Anyone with templates | Admin only |
| Publish from a draft | Anyone with templates | Admin only |
| Restore an earlier version | Anyone with templates | Admin only |
| Start or promote a staged rollout | Anyone with templates | Admin only |
| Schedule a publish for later | Anyone with templates | Admin only |
| Abandon a running rollout | Anyone with templates | Anyone with templates |
Every route that changes what users receive is gated, rather than the publish button alone. A gate that covered publishing by itself would be decorative, because restoring a version reaches users just as directly.
Abandoning a staged rollout is the deliberate exception and is never gated. Whoever notices a problem has to be able to stop it, and stopping a rollout puts users back on the body that was already approved.
The review cycle
Section titled “The review cycle”An editor saves a draft and submits it for review. The template then carries an Awaiting review badge on the Templates page, naming who submitted it, and the editor shows when.
An admin opens it, reads the line diff against the live body, and either publishes it or sends it back. A designer template has no diff, so there the review is the canvas and the preview.
Both editors carry the whole cycle. Approval would otherwise be a trap rather than a control for designer templates: they only ever open in the designer, so an editor refused a publish with no draft to submit would have had nowhere to go.
Approving is publishing. There is no third state where a draft has been approved but is not yet live, because such a body would be a fourth kind of content alongside the live one, the draft and a canary, with no answer for what should happen when the live body changes underneath it.
Sending it back needs a reason. A rejection with no note is a dead end for whoever has to act on it, so Sigil requires one and keeps up to 1,000 characters of it.
The body is left exactly as it was when it is sent back. A rejection is a request for changes rather than an undo, so the editor reopens the same work with the note attached to it.
Submitting a draft for review
Section titled “Submitting a draft for review”With approval on, anybody who is not an Admin sees Submit for review in place of Publish.
- Open the template in its editor from Templates.
- Make your changes.
- Choose Submit for review. It saves what is in the editor as the draft and submits it in one step.
A message confirms an admin will publish it or send it back, and the button reads Awaiting review until they do. On the Templates page the template carries an Awaiting review badge. Saving another draft takes it out of the queue.
Approving a draft
Section titled “Approving a draft”You need to be an Admin.
- Open Templates in the portal sidebar and find the template with the Awaiting review badge.
- Choose Edit or Design to open it.
- Review the submitted version. In the HTML editor, a Waiting for your review card sits at the top and Diff compares it against the live body. In the designer, a notice names who submitted it.
- Choose Approve and publish.
- In the HTML editor, choose Approve and publish in the confirmation.
A message confirms it is approved and published.
Sending a draft back
Section titled “Sending a draft back”- Open the template from Templates, as for approving.
- Choose Send back.
- Under What needs changing?, type your note.
- Choose Send back.
The author sees the note in the editor, marked as sent back by you, the next time they open the template.
The author hears the outcome
Section titled “The author hears the outcome”Whoever submitted the draft is emailed when it is decided, so they do not have to keep opening the template to find out.
| Outcome | The email |
|---|---|
| Approved and published | “Your changes to” the template’s name “are live”. Names who approved it |
| Sent back | “Your changes to” the template’s name “were sent back”. Names who sent it back, quotes your note in full, and says the draft is kept so they can change it and submit again |
An approver who reviews their own draft is not sent either one, since they already know.
The approvers hear about it
Section titled “The approvers hear about it”Submitting a draft emails every other user with the Admin role in the organisation. The message names who submitted it and which template, and its button opens the library to review it.
The mail also carries a picture of the proposed signature, and of the live one above it where there is one, so an approver can judge an ordinary change from their phone without signing in. The pictures ride inside the message rather than being linked, because Outlook hides remote images from senders it does not know and an approval email that arrived as two empty boxes would be worse than the text it replaced. They show the sample person, so the message discloses no colleague’s details to whoever reads it.
The submitter is not sent a copy, and an organisation with a single administrator who submitted the draft themselves has nobody to notify, so nothing is sent.
The mail is built shortly after the submit rather than as part of it, and the draft is read again at that point. A draft withdrawn, published or edited in the meantime sends nothing rather than a picture of a body that no longer exists.
The queue in the portal and the next health digest both still list a submitted draft, so the mail is a prompt rather than the only place a review appears.
Editing resets the review
Section titled “Editing resets the review”Saving a draft clears its review state. A submitted draft drops out of the queue, and a rejection note stops being displayed.
This is deliberate. A body that has changed is not the body a reviewer was looking at, and a note that describes text somebody has since rewritten is worse than no note. The change log keeps both permanently, so nothing is lost.
Resubmitting a rejected draft clears the note for the same reason: a resubmission answers whatever the last reviewer objected to.
Submitting without approval turned on
Section titled “Submitting without approval turned on”The server accepts a submitted draft whether or not the switch is on, so a script can ask for a second pair of eyes on one significant change without the organisation imposing a review step on every typo fix.
The portal does not offer it. Both editors show Submit for review only to somebody who cannot publish, and with the switch off everybody with templates can, so the button never appears there while approval is off.
Approving your own work
Section titled “Approving your own work”An admin can publish a draft they submitted themselves.
A strict four-eyes rule would lock an organisation with a single admin out of its own signatures, which is a worse failure than the one it prevents. Sigil permits self-approval and records it instead: the change log entry names who submitted, who approved, and whether those were the same person.
That makes the control auditable, which is what an auditor is actually asking for. If your own policy requires two people, the log is where you evidence it.
What an editor sees when they are refused
Section titled “What an editor sees when they are refused”A publish attempt that approval blocks returns a message saying the organisation requires publishing to be approved, and telling them to save the change as a draft and submit it.
The portal hides the buttons an editor cannot use, but the server checks independently, so the refusal is the control rather than the hidden button.
What is recorded
Section titled “What is recorded”Submitting and sending back each write a change log entry under the Approval action, with the transition and the rejection note.
Approving is a publish, so it appears as a Published entry. A publish that came from a submitted draft records the submitter, the approver and whether the two were the same person.
Partner-managed organisations
Section titled “Partner-managed organisations”A partner Owner or Admin working inside a managed client can change the client’s approval switch. A partner Technician cannot.
Partner Owner and Admin staff working inside the client also count as approvers: they can publish, approve and send back. A partner Technician works as an Editor does and submits for review. The approval email goes only to the client’s own Admins, never to partner staff, so for a client with no Admin of its own, watch for the Awaiting review badge on the Templates page instead.
Turning off a client’s publish approval is exactly the kind of change the control exists to prevent, so it sits with the partner roles that also manage access rather than with the role that does the signature work. See partner roles.
When it is the wrong tool
Section titled “When it is the wrong tool”Approval slows every publish down by however long it takes an admin to look. That is the point of it, and it is a poor trade for an organisation where the same person would be doing both halves.
If what you actually want is confidence that a large change works before everyone gets it, a staged rollout measures that directly and needs nobody to be watching.
The two compose. A rollout can be started by an approver, and it will still stop itself if the new body starts failing.
