feat(newos4): on-demand worker labels and a server-canonical predictor store (v4.6.70)

Two decisions recorded by the operator on 2026-10-23, implemented.

6 — spawn-on-demand is the design, so stop reporting it as a defect:
- OrchestratorModule counts delegations inside delegateTask()
- status prints "Mode: on-demand — N/M worker(s) available, X delegation(s) this
  session" and drops the "spawning is not dispatching" warning
- per-agent lines read "available, no delegations yet (spawned ...)" / "N
  delegation(s), last ..."

10 — per-user, server file canonical, browser VFS a TTL'd read-through cache:
- store = the user's own WebDAV space (<space>/_data/predictor.json), so identity
  comes from credentials the app already has; no new endpoint, no unauthenticated
  write surface
- 5 min TTL: cache wins while fresh, server wins when it differs; a banner naming
  both counts prints only on divergence
- write-through on save (debounced 5 s) with a MKCOL retry when the collection does
  not exist yet; store mode is printed by predictor-status
- the cache write is guarded and verified — a missing /system/metrics made writeFile
  fail silently, which is how the first version lost a pull

Verified live on v4.6.70 with a stubbed transport: pull applies 5 pts/5 paired and
writes a stamped cache, the divergence banner fires once and not on the second
(equal) pull, write-through PUT is issued, status reads "Store: server". A real
WebDAV round-trip still needs the user's WebDAV config in-app; the agando space has
no _data collection yet, so GET returns 404 -> "server (empty)".

1 files changed, 12 insertions, 6 deletions

diff --git a/modules/core/module-orchestrator.js b/modules/core/module-orchestrator.js
index d03f7eb..b3b76cd 100644
--- a/modules/core/module-orchestrator.js
+++ b/modules/core/module-orchestrator.js
@@ -25,6 +25,9 @@ class OrchestratorModule {
         this._a2aPollTimer = null;
         this._heartbeatInterval = null;
         this._a2aMessages = {};
+        // v4.6.68: on-demand is the design (decision 2026-10-23). Delegations — not
+        // "agents ever scheduled" — is the number that means something.
+        this._delegations = 0;
         this._restartBackoffs = {};
         this._eventBus = {};
         this._idleSweepInterval = null;
@@ -1077,6 +1080,7 @@ class OrchestratorModule {
 
     async delegateTask(task, opts = {}) {
         if (!task) return this.terminal.print('Usage: orchestrate <task description>', 'error');
+        this._delegations = (this._delegations || 0) + 1;
 
         // Parallel fan-out: decompose and run sub-tasks concurrently
         if (opts.parallel) {
@@ -1423,7 +1427,7 @@ class OrchestratorModule {
         // v4.6.63: say what the state MEANS. A parked agent with 0 assignments and a parked
         // agent between tasks looked identical from outside (packet §6).
         const agentsArr = Object.values(this._agents);
-        const everScheduled = agentsArr.filter(a => (a.tasksAssigned || 0) > 0).length;
+        const available = agentsArr.filter(a => a.status === 'idle' || a.status === 'sleeping').length;
         const describe = (a) => {
             const n = a.tasksAssigned || 0;
             const since = a.createdAt ? String(a.createdAt).slice(0, 16).replace('T', ' ') : '?';
@@ -1436,19 +1440,21 @@ class OrchestratorModule {
             }
             if (n === 0) {
                 return (a.status === 'sleeping' ? 'parked' : a.status) +
-                    ` — NEVER SCHEDULED (0 tasks since ${since})`;
+                    ` — available, no delegations yet (spawned ${since})`;
             }
             return (a.status === 'sleeping' ? 'parked' : a.status) +
-                ` — waiting for work (${n} task(s), last ${last || '?'})`;
+                ` — ${n} delegation(s), last ${last || '?'}`;
         };
         agentsArr.forEach(a => {
             const modelInfo = a.model ? ` [model: ${a.model}]` : '';
             const backoff = a.nextRestartAt && a.nextRestartAt > Date.now() ? ` [restart in ${Math.round((a.nextRestartAt - Date.now()) / 1000)}s]` : '';
             this.terminal.print(`  ${a.id} [${a.role}] — ${describe(a)}${modelInfo}${backoff}`, 'info');
         });
-        this.terminal.print(`  Dispatch: ${everScheduled}/${total} agent(s) have ever been given a task`, everScheduled === 0 && total > 0 ? 'warning' : 'info');
-        if (everScheduled === 0 && total > 0) {
-            this.terminal.print('  ⚠️ No agent has ever been scheduled — spawning is not dispatching. Use orchestrate <task> or orchestrator-spawn <role> <task>.', 'warning');
+        // v4.6.68: on-demand by design — a worker with nothing to do is healthy, not broken.
+        // The number to watch is how many delegations the AI actually made.
+        this.terminal.print(`  Mode: on-demand — ${available}/${total} worker(s) available, ${this._delegations || 0} delegation(s) this session`, 'info');
+        if (total > 0 && (this._delegations || 0) === 0) {
+            this.terminal.print('  Workers wait for delegate_task — spawn-on-demand, no background dispatch queue by design.', 'text-muted');
         }
     }