openprose__prose/packages/std/ops/prose-author.incident-response.test.prose.md

name, kind, subject

name	kind	subject
test-prose-author-incident-response	test	prose-author

Test Prose Author Incident Response

Fixtures

• request: | incident response workflow

  input incident_id, severity, affected_service

  triage: gather current alerts, recent deploys, active incidents, owner rotation, and customer impact for affected_service

  if severity is sev1 or sev2: page incident commander and service owner create incident channel and status-page draft else: notify service owner in team channel

  loop until mitigated or max 4 cycles: investigator session: inspect telemetry and logs, propose likely cause and confidence mitigator session: propose lowest-risk mitigation, including rollback or feature flag options reviewer session: approve mitigation only if blast radius and rollback plan are clear if reviewer approves: execute mitigation checklist and watch metrics for 10 minutes else: feed reviewer notes into next cycle

  after mitigation: publish customer/status update if external impact create postmortem issue with timeline, suspected cause, followups, and owners

  return incident_summary, mitigation_decision, postmortem_issue

Expects

• source_package: prefers folder output because the workflow has branching, operational side effects, a bounded mitigation loop, and several specialist services
• source_package: maps incident_id, severity, and affected_service into ### Requires
• source_package: preserves the return outputs incident_summary, mitigation_decision, and postmortem_issue
• source_package: includes explicit if/else ProseScript for the severity branch
• source_package: includes a bounded mitigation loop preserving max 4 cycles
• source_package: carries reviewer notes into the next mitigation cycle
• source_package: gates mitigation execution behind reviewer approval
• authoring_notes: records that operational side effects are represented as generated declarations and future runtime actions, not executed during authoring
• source_package: declares degraded outputs or errors for exhausted mitigation attempts
• lint_report: has status pass and no blocking findings

Expects Not

• authoring_notes: claims PagerDuty, Slack, status pages, issue trackers, deployment systems, or feature flag systems were contacted during authoring
• source_package: executes mitigation before reviewer approval
• source_package: treats paging, channel creation, status updates, or issue creation as ordinary text with no side-effect boundary
• source_package: silently returns success when mitigation was not approved or not confirmed