the related-issue dump is the useful bit. agents keep reopening stuff that’s already fixed two tickets over. if this stops that loop i’d actually run it before starting a session.

Any plans to support Linear?

engagement points the wrong way for an agent factory: the most-discussed items are usually the ones needing the most human judgment, so ranking by it queues agent hours behind review slots. rank by 'can this run end to end unattended' and throughput stops being human-gated.

This feels like an important layer for multi-agent development. Before assigning work, understanding related issues, existing fixes and conflicting PRs can matter as much as the coding task itself.

This is where AI coding gets really interesting: the bottleneck shifts from “what should we build?” to “what should we build next?” Turning the backlog into a continuously ranked set of opportunities could be a huge unlock for agent-driven teams.

vercel never disappoints 🙌🏻

So cool

that's insane