Skip to content

Where cross-links are delivered

Approving a cross-link decides that it should exist. This page is about where the change turns up once you've approved it — the repository, the ticket, the endpoint your team actually watches.

Different teams work in different places, so Kenaptic delivers to several. What never changes is who decided: the approval always happens in Kenaptic, by a signed-in person. Everything below is delivery, not a second gate.


Changes land in your repository

For docs-as-code content — a documentation site, handbook or blog built from Markdown in a repository — Kenaptic opens a change the way one of your own engineers would.

Host How the change arrives
GitHub A pull request
GitLab A merge request
Bitbucket A pull request
Azure DevOps A pull request

You don't pick the host from a menu. Kenaptic works it out from the repository address you gave it, so it's impossible to select "GitHub" for a Bitbucket repository and then spend an afternoon puzzling over an error from the wrong service.

Branch or fork — Kenaptic checks, rather than assuming

For your own repository, a branch pushed straight to it is what you want: the proposal shows up in the repository your reviewers already watch. For a repository you don't own, a fork is the only correct route.

Kenaptic tests what the credential you supplied is actually allowed to do, and takes the direct-branch route only when it can. A capability check beats a setting here, because a wrong setting fails at the very end — after the content has been read, the link written and the commit made — which is the most annoying possible moment to discover it.

Whichever route it takes, the change itself is identical: the same marked, reversible link block described in How linking stays safe. Re-running doesn't open a second pull request for the same page — it updates the one already open.

Content with no repository behind it

Plenty of good content isn't in git. A knowledge base, a CMS-hosted blog, a community platform. Kenaptic writes to those through the platform's own official interface, using a credential you create and control, and the same rules apply: marked, reversible, and only ever after approval.

On community platforms Kenaptic adds a reply pointing at the related page. It never edits somebody's post — those words belong to the person who wrote them.


Records in your ticketing system

Some teams need the proposal to appear in the system where their work is tracked and audited, not only in Kenaptic. So Kenaptic can also file a record describing each proposed cross-link.

Destination Status What arrives
Jira Available An issue in the project you choose, updated by comment as things change
Generic webhook Available A structured JSON payload to any endpoint you run
ServiceNow Groundwork in place A record on the table you nominate

Each record carries the two pages, the relationship type, and a deep link straight back to the item in Kenaptic. Delivery is idempotent — if the same proposal is delivered twice, you get one ticket that gets updated, not a duplicate.

Delivery is one-way, on purpose

Kenaptic does not read a status back. Closing the Jira ticket does not approve the link, and resolving the ServiceNow record does not publish it.

This is a deliberate line, not a missing feature. Approving a cross-link is the moment a change to your live content becomes real, and Kenaptic will only accept that instruction from an authenticated person in the app, with 2FA behind it. Anything else means the audit trail records a workflow transition rather than a human — and "the ticket was closed" is not an answer to "who approved this, and when?".

Use the ticket for the things tickets are good at: scheduling, assignment, visibility, and the record your process needs. Use Kenaptic for the decision.

What 'groundwork in place' means for ServiceNow

The ServiceNow path is built and will file a record, but every ServiceNow instance is customised — different tables, different mandatory fields, different approval flows. We've left a clear seam for those extra fields rather than guessing at a configuration nobody asked for. If you run ServiceNow and want this finished properly, talk to us: we'd rather build it against your actual instance than against our imagination of it.


Before you publish: check there's a way through

A cross-link with nowhere to go is the worst kind of failure, because everything looks fine right up to the end. Kenaptic therefore checks each destination before you publish, and shows a warning marker in the preview list against any source that has no usable write path — no credential, or no repository destination configured.

Hover it and Kenaptic tells you exactly what's missing. Fix it in Domains, or approve the link anyway and leave it queued until the path exists — but you'll know either way, rather than finding out from a publish that quietly did nothing.


Choosing where things go

  • Your content is in git and your team reviews pull requests. Use the repository path alone. It's the most direct route and the reviewers are already there.
  • Your content is in git, but a change to published docs must be tracked in Jira. Use both: the pull request carries the change, the Jira issue carries the record.
  • Your content isn't in git. Use the platform's own interface, and add a ticketing destination if your process needs one.
  • You have an internal workflow tool of your own. Use the generic webhook and route the payload wherever you like.

Next