The first example is expanded. The gallery shows the same pattern across regulated teams: draft the issue, fill schema-valid fields, reject unsupported values, and leave an audit trail.
Customers cannot send payments to Thailand after yesterday's release. Error code 500 appears in payout flow. Started around 08:00 UTC.
Drafted summary, acceptance criteria, and triage fields from the team's own runbooks and prior incidents.
Fields: severity, region, owner. Rejection: unsupported sanctions category. Sources: runbook + prior incident. Audit: draft.created. Example manual work avoided: 12 minutes.
Fields: data class, approver group, rollback task. Rejection: ungrounded root cause. Sources: access policy. Audit: reviewer required. Example manual work avoided: 15 minutes.
Fields: impact, environment, review gate. Rejection: missing owner. Sources: deployment guide. Audit: offline path. Example manual work avoided: 10 minutes.
Fields: acceptance criteria, audit owner, enclave label. Rejection: external endpoint. Sources: package runbook. Audit: no vendor egress. Example manual work avoided: 18 minutes.
Fields: component, owner, stale reason. Rejection: unknown priority. Sources: issue history. Audit: review-only changeset. Example manual work avoided: 7 minutes per issue.
Fields: control owner, evidence type, due date. Rejection: unsupported control id. Sources: policy pack. Audit: export ready. Example manual work avoided: 20 minutes.
Why this matters: the trust signal is not that every field is filled. It is that unsupported values are visibly rejected and never written to Jira. Manual-work estimates above are examples for pilot planning, not measured public claims.
The test drive takes the payments example through draft, validation, rejection, audit, and egress proof.