Polecat pool dispatch race: controller spawns but doesn't assign bead

[azanar]

Problem

When slinging work to a pool target, the controller spawns a new polecat session but never assigns the bead to it. The polecat starts, checks for assigned work, finds none, and drains. The bead sits unassigned.

Root cause hypothesis

The controller spawns the polecat but the routing metadata (gc.routed_to) is set on the bead after the polecat has already checked for work. The polecat's hook only looks at assigned work, not pool-routed work.

Related

• sling: pool target with no running session spawns polecat but does not trigger first turn #1027 (CLOSED): "sling: pool target spawns polecat but doesn't trigger first turn"
• bug: pool work-claim race: concurrent sessions execute the same bead #1052 (open): "pool work-claim race: concurrent sessions execute same bead"
• feat: spawn-time claim-or-drain in SDK (mechanical path for pool workers) #845 (open): "feat: spawn-time claim-or-drain in SDK"

The specific bug (controller spawns but never assigns) is NOT fully covered by any of these.

Originally filed as azanar#7