Slack On-Call Handoff Workflow Prompt

You are a senior SRE who has standardized on-call handoffs for a team running 24/7 rotations with a measurable drop in incidents-missed-because-of-handoff after rollout.

I will provide:
- Our scheduling tool (PagerDuty / OpsGenie / Grafana OnCall)
- Where active incidents live (the same tool, or FireHydrant / Incident.io)
- Ticket system (Jira / Linear / GitHub Issues)
- Change tracking (deploys, infra changes — ArgoCD, Spinnaker, Terraform Cloud, GitHub)
- SLO source (Sloth, Nobl9, Datadog, custom)

Your job:

1. **Handoff message structure** — every section's purpose, position, and skip-ability:
   - **Header**: outgoing → incoming, time, primary + secondary, link to scheduler
   - **Active incidents**: SEV, age, owner, last update, link
   - **Recent changes (last 24h)**: deploys with revert links, infra PRs, feature flags toggled
   - **Open follow-ups from prior on-call**: action items not yet closed
   - **Hot tickets to monitor**: bug reports with high reproduction rate
   - **SLO burn**: services in fast or slow burn windows
   - **Watch items**: scheduled maintenance, known fragile components, ongoing migrations
   - **Footer**: how to ack, what to do if you weren't paged but should have been

2. **Data sources & queries** — for each section, the exact API call or query against your tools. Show the field-mapping needed.

3. **Generation cadence** — daily at handoff time (configurable per team — Pacific morning is European afternoon), plus on-demand `/handoff` slash command.

4. **Review & approval** — generated draft posts to a private thread; outgoing engineer can edit before publishing to the team channel. Default-publish after N minutes if no edits.

5. **Storage & search** — published handoffs land in a searchable index (Loki / Elasticsearch / Notion). Easy to answer "what was happening last Tuesday at 06:00?"

6. **Anti-patterns to avoid** — generated wall of unfiltered links, no priority signal, no clear "thing to do RIGHT NOW", buried watch-items, no continuity between days.

7. **Validation** — after 4 weeks: measure (a) % handoffs with active SEV1/2 successfully picked up without re-paging, (b) MTTR on incidents that span shift changes, (c) reduction in "I missed that you were dealing with X" messages.

Output as: (a) handoff template in Block Kit JSON, (b) per-section data-source mapping table, (c) generation pipeline (cron + script structure), (d) review/approval UX, (e) rollout plan with measurement.

Bias toward: actionable items at the top, scannable in 60 seconds, automated facts but human-curated priorities.