Use casesStatus updates

The on-call handoff note, written before the shift changes

Mitra can write your on-call handoff note: every Monday at 9am, an agent reads the week's PagerDuty pages, matches them to the fixes merged in GitHub, and publishes a Notion note covering what paged, what caused it and what is still unresolved. The summary also goes to Slack.

Runs: Mondays at 9am
Used by: Engineering · Ops / Chief of Staff

The prompt

Every Monday at 9am, go through the PagerDuty incidents from the past week, pair each one with the pull request that fixed it, and write a handoff note in Notion listing what paged, the cause and what is still open. Post a short version to #on-call.

PagerDuty

GitHub

Notion

Slack

Build this agent

200 credits to start · no credit card

What the agent actually does.

Read the week's pages
The agent pulls every PagerDuty incident from the past seven days, with who was paged, when, and how long each took to resolve.
Pair each page with its fix
Incidents are matched to the pull requests and commits that closed them, so the note says what was actually changed.
Separate open from closed
Anything without a resolution or a merged fix goes into an open list, with how long it has been sitting there.
Write the note in Notion
One dated page per shift: what paged, what caused it, what changed, and what the next person inherits.
Post the summary to Slack
A short version lands in the on-call channel as the shift starts, linking to the full note.

Make it yours.

For teams who hand on-call between people every week, where the handoff today is a verbal summary from whoever remembers the worst night. Change any of this by saying so — there is nothing to rewire.

Send the note as a direct message to the incoming on-call instead of posting in the channel.
Run at the end of the shift rather than the start, while the week is still fresh.
Add a count of pages per service, so the noisiest service is obvious at a glance.

Questions, answered.

What should an on-call handoff note include?

A good handoff answers three things, and Mitra writes all three: what paged during the shift, what caused each page, and which threads the next person inherits. The first two come from PagerDuty and GitHub. Anything with no fix merged lands in the open list.

How do I hand off an incident that isn't fixed yet?

Mitra gives unresolved pages their own list in the handoff note, with how long each has been open. The agent looks for an ending — a resolved incident in PagerDuty, a merged fix in GitHub, a closed issue — and anything without one stays on that list. The next on-call inherits a short read.