incident.io process hygiene board for weekly reviews

Build me an internal app that I open before our weekly incident review, so we can see whether we actually followed our own incident process rather than just how many incidents we had. It reads from incident.io and can post to Slack using the Slack Bot integration. Everything is read fresh from incident.io when I load the board. The only things the app stores itself are my configuration and the agendas it generates.

The main view is a hygiene board. Each row is an incident from the selected period, pulled with List Incidents, showing its reference, name, severity, incident type, current status and when it was reported. Every column after that is a process check scored green, amber or red: severity set, lead assigned, required custom fields filled in, status updates posted often enough for that severity, resolved timestamp recorded, and follow-up actions logged. Clicking any cell explains why it scored the way it did, so nobody has to take the colour on faith.

Build the checks from these reads. Severity set compares the incident's severity against the list from List Severities and scores red when nothing is assigned. Lead assigned uses List Incident Roles to find the built-in lead role and checks whether the incident has anyone in that seat, scoring red when it is empty. Required custom fields uses List Custom Fields for the definitions plus the incident's own field values, scoring green when every field I marked required is filled, amber when only some are, and red when none are. Resolved timestamp uses List Incident Timestamps to find the resolved timestamp definition and checks whether the incident has a value for it, scoring red for a closed incident with no resolved time and showing not applicable while the incident is still open. Follow-up actions logged uses List Actions and List Follow-ups filtered to the incident, scoring green when at least one is logged, amber when they are logged but nobody owns them, and red when there are none at all.

The update cadence check is the one that needs real computing. Pull the incident's status updates with List Incident Updates, sort them by time, and measure the gap between each consecutive update, plus the gap from when the incident was reported to the first update and from the last update to resolution. Compare the largest gap against a target I configure per severity, for example every 30 minutes for a SEV1 and every 2 hours for a SEV3. Score green when every gap is within target, amber when the worst gap is up to a tolerance multiple of the target that I can set, and red beyond that or when there were no updates at all. Show the worst gap in the cell so I can see how badly it slipped rather than just that it did.

Let me filter the board by severity, by incident type from List Incident Types, and by date range, defaulting to the last seven days. Above the board show a summary strip: how many incidents in the period are clean on every check, and a count per check of how many incidents failed it, so the gaps that keep repeating are obvious at a glance.

Clicking a row opens a detail panel with the full incident from Show Incident, a timeline of its status updates with the gaps marked, and its outstanding actions and follow-ups. From the row or the panel I can fix gaps in place: backfill a missing severity, assign a lead, or fill in a required custom field using Edit Incident, applied straight from the board and reflected in the row without a full reload. I can also nudge the incident lead with the Slack Bot Send a Message action, which prefills a short message naming the incident and the specific gaps and lets me edit it before it goes.

Add a Prep the review button that runs a background agent over the currently selected period and filters. The agent reads each incident in scope with Show Incident, its status updates with List Incident Updates, and its outstanding work with List Actions and List Follow-ups, then writes a meeting agenda back into the app: which incidents need real discussion and why, which are clean enough to skip, and which process gaps keep repeating across the period. Show the agenda in its own view with the period it covers and when it ran, keep past agendas so we can tell whether the same gaps keep coming back week after week, and let me copy it or post it to a Slack channel with Send a Message.

Give me a settings view for the parts that are opinions rather than data: the update cadence target for each severity, the tolerance multiple that separates amber from red, and which custom fields count as required. Store those in the app so they persist between sessions, and seed the severity list from List Severities so I am configuring against our real severities. Every team sets its own communication expectations, so none of these thresholds should be hardcoded.