/.github/loop/PRIORITIES.md

Priorities

The ranked work queue for the autonomous improvement loop. The architecture-review pass (the architect) owns this file: each run it turns the roadmap plus an internal scan (gaps in the services → agents → workflows lifecycle, API coherence, drift, tech debt, test and DX friction) into a single ordered list — highest-value first — and links each item to a tracking issue. The hourly continuous-improvement pass works the top item whose issue is still open. So the architect decides what, and the increment loop builds it.

Reading / editing. An item is done when its linked issue closes (the increment that builds it adds Closes #<issue>). Roadmap phase (Now → Next → Later) is the primary ordering; internal findings are interleaved by value, not kept in a separate list. The human can reorder this list — or the issues — at any time to redirect the loop; direction always wins.

Off-limits to the loop (the architect proposes these as notes, never as queue items the loop can auto-merge): brand/positioning copy, breaking public-API changes, architectural rewrites. Those go to the human.

Work queue (ranked)

1. Handle incomplete AtlasCloud plan-delegate task list tool calls (#4606) — Highest-value open Now-phase hardening item after #4599 shipped in #4608: the 0→hero plan/delegate path already created a task, then failed on an incomplete repaired task_TaskService_List tool call. Fix this first because partial side effects in the canonical services → agents → workflows demo directly undermine the adoption story even when the mock harness is green.
2. Handle incomplete AtlasCloud agent-flow workspace tool calls (#4595) — Same provider-repair seam in the event-driven agent-flow path: the expected workspace/notification side effects occurred once, duplicate side effects were suppressed, then the flow still reported failure on an incomplete repaired workspace tool call. Keep it adjacent to #4606 so the runtime handles malformed repaired tool calls consistently across plan/delegate and flow dispatch.
3. Make AtlasCloud universe A2A reachability probe deterministic (#4504) — The remaining live AtlasCloud interop seam is the universe A2A reachability probe intermittently timing out after the checkout flow succeeds. Keep it behind the repaired-tool-call failures because it is side-effect safe and isolated to reachability/probe timing rather than the core 0→hero side-effect path.

Seeded by Claude Code from the roadmap + open issues; thereafter maintained by the architecture-review pass.