Incidents

What broke, who's on call, what happened next.

Incidents move across a three-column board with a severity chip on each card; the write-up lives on the record, so the post-mortem starts already half-written.

The second tab is the on-call schedule: who holds the pager and when. Agents that watch alerts can file incidents here and read who is on call before they escalate.

What it does

Open / investigating / resolved board with severity chips.
The write-up lives on the incident record.
On-call shifts with dates, assigned to employees.
Agents file incidents and check the rota over MCP.

What it stores

As ordinary objects in your workspace memory — readable by your AI over MCP, browsable like any memory.

incident — Something that broke — status (open|investigating|resolved), severity (low|medium|high|critical), opened_on; the write-up is the description.
oncall_shift — One on-call rotation — from_date to to_date, held by an employee.
employee — A person employed by the company — job_title, start_date, salary, status (active|left).