An Orkeste agent reads your Linear issues on the schedule you set and writes the sprint update: what shipped, what is in progress, what is blocked and what is at risk. It posts the update to Slack once the engineering lead has approved it.
The problem
Status updates are written from memory at the end of the week, or replaced by a meeting. Stakeholders ask the same question in five places.
How the workflow runs
Runs: Every day or every week at the hour you set.
Search Linear: The agent searches Linear for the team's issues in the current cycle, up to 20 per search.
Group the work: It sorts issues into done, in progress, blocked and at risk.
Write the update: It writes a short update that people outside engineering can read.
Approve: The update waits in Validations for the engineering lead.
Post: The approved update is posted to Slack, and can also be saved on a Confluence page.
The update is shared only after the engineering lead approves it.
Who does what
Sprint reporter: Turns the cycle's issues into a readable update.
Engineering lead (human): Approves the update before it is shared.
What this workflow does not do
A Linear search returns up to 20 issues; for large teams, use one agent per team.