Issue triage

Linear holds issues, bugs, and feature requests for software teams. A Beam agent watches for new issues, reads the title, description, and labels, and applies your rule to set priority, add labels, and assign the right owner. When an issue is unclear, duplicates an open one, or lacks the detail to act on, the agent flags it for a lead. A person makes the call on vague reports and decides which issues belong in the current cycle for the team. The agent leaves a brief note on each issue explaining the labels and priority it chose so the lead can check its reasoning fast.

Cycle planning

Linear organizes work into cycles so teams commit to a set scope. A Beam agent reads issue estimates and status across a cycle, applies your rule to flag overcommitment or stalled work, and updates the cycle view. When a cycle is trending past its capacity, the agent notifies the lead with the issues at risk. People decide what to defer, what to pull in, and whether an estimate needs revisiting before the next planning session begins in earnest. The agent refreshes the cycle view as issues close so the team always sees how much scope is still open.

Status updates

Linear advances issues through states as work progresses. A Beam agent reads updates from linked branches, comments, and pull requests, applies your rule to advance an issue's state, and notifies the assignee of anything waiting on them. When a signal is ambiguous, such as a branch merged without a clear issue link, the agent asks a person rather than guessing. Humans confirm when work is truly done and decide how to handle issues that stall mid cycle. The agent keeps the assignee posted on each move so the board reflects real progress without anyone updating it by hand.