github-triage

Automating GitHub issue triage uses an AI-assisted, label-based state machine to read issue content, suggest state transitions like unlabeled to needs-triage, and generate actionable triage briefs to reduce manual triage time.

A label-based state machine for GitHub issue triage operates across states like needs-info, ready-for-agent, ready-for-human, and wontfix, reading issue content and prior notes to determine the appropriate workflow and next actions.

Generating agent briefs from GitHub issue content involves reading the issue body, comments, labels, and internal references to synthesize a precise triage brief that surfaces context for AFK agents or human handlers.

Determining if a GitHub issue is out of scope during triage involves reading the issue content against a dedicated out-of-scope knowledge base to decide whether to apply the wontfix label and generate appropriate comments.

AI can suggest state transitions for both bug and enhancement issues on GitHub by reading issue content, labels, and prior triage notes to guide actions across needs-triage, ready-for-agent, and ready-for-human states.

Preparing a GitHub issue for an AI agent involves synthesizing the full issue content, prior triage notes, and internal references into an actionable agent brief that transitions the issue to the ready-for-agent state.