darkrun-composite

name	darkrun-composite
description	Create a composite Run combining stations from multiple factories with sync points

Create a composite Run that coordinates work across two or more factories in parallel, with sync points where one factory's line must wait on another's.

Process

1. Gather from the user:

   • The work description
   • The factory selection — 2+ required (use darkrun_factory_list for what's available; one factory = a normal darkrun_run_new)
   • The station subset per factory (darkrun_factory_detail shows each line; empty = the full line)
   • Where the factories must synchronize

2. Suggest sync points from the artifact chains: when one factory's station consumes what another's produces, that's a wait/then edge. Handles are factory:station.

3. Create it:

darkrun_run_composite {
  slug, title,
  parts: [
    { factory: "software", stations: ["build", "prove"] },
    { factory: "legal",    stations: [] }
  ],
  sync: [
    { wait: ["software:prove"], then: ["legal:shape"] }
  ]
}

4. Report the created Run with the part/station overview and the sync map.

Coordinating a composite Run

A composite Run is not single-walkable — darkrun_tick surfaces the topology instead of walking it. Drive each part as its own Run (darkrun_run_new per part, scoped with size/station subset), honor the sync points (don't start a then part before its wait handles complete), and record progress on the ledger as parts move:

darkrun_composite_stamp { slug, handle: "software:build", note: "completed" }

The ledger (composite_state) is the coordination record the operator reads to see where the composite stands.