Recert

Public pages for the Recert app: getting started, privacy policy, security, support

View the Project on GitHub ezekiel06/recert-site

SOC 2 and ISO 27001 user access reviews for Jira and Confluence

If your company is going through SOC 2 or ISO 27001, at some point the auditor asks for evidence that access to your systems is reviewed periodically. ISO 27001:2022 says it directly in Annex A control 5.18, Access rights; SOC 2 auditors test it under the logical access criteria (CC6). Jira and Confluence usually end up in scope because that is where product plans, customer tickets and internal documentation live.

This page is about what that evidence has to contain, and where Jira and Confluence make it hard.

What the evidence needs to show

For each review period, and for each Jira project and Confluence space in scope:

  1. A dated list of who had access. Taken at a known point in time, and complete: people who get in through groups and permission schemes count, not only the ones listed by name.
  2. Why each person had access. The group, role or scheme grant. Without it, a reviewer cannot act on “remove”, and the auditor cannot tell whether the list is complete.
  3. Who reviewed it. Normally the owner of the project or space, not one admin for everything.
  4. A decision per person. Keep, revoke, or an accepted exception, with a reason for the last two.
  5. Proof that revokes happened. A ticket, a change record, or a later snapshot showing the person gone.
  6. Sign-off. That the review was completed, by whom, and when.

Where Jira and Confluence make this hard

Common reasons this evidence fails

What Recert does

Recert produces the evidence above for Jira projects and Confluence spaces in one review. It snapshots effective access, including access through groups, roles, permission schemes and license classes, leaving app accounts out. It routes one review to each project lead or space administrator, records keep / revoke / exception decisions with reasons, tracks sign-off, and exports the result as CSV files attached to a Jira issue. It runs entirely on Atlassian’s infrastructure; no data leaves your site.

See the getting-started guide.