What problem does it solve? Maintainers lose track of incoming bugs and feature requests, and issues handed to autonomous agents often lack durable, testable specifications. This Skill moves issues through a consistent triage state machine and produces structured agent briefs that stay valid as the codebase changes. ## Core Features & Use Cases - State Machine Triage: Classify issues with category roles (bug, enhancement) and state roles (needs-triage, needs-info, ready-for-agent, ready-for-human, wontfix), with conflict detection and maintainer overrides. - Attention Buckets: Query the issue tracker and present unlabeled, needs-triage, and needs-info issues oldest-first, excluding meta issues. - Agent Brief Generation: Write durable, behavioral specifications with acceptance criteria and explicit out-of-scope boundaries for AFK agents, following the AGENT-BRIEF.md reference. - Rejection Records: Record rejected enhancements in docs/adr/ (ArkEnv override) instead of a root .out-of-scope/ directory, and surface prior rejections during triage. - Use Case: A maintainer says "Let's look at #42" — the Skill reads the issue, checks ADRs for prior rejections, attempts reproduction for bugs, recommends a category and state, then posts an agent brief or needs-info notes. ## Quick Start Ask your AI assistant to triage issue #42 and move it to ready-for-agent with an agent brief.