feat: probe/teach agent split — explicit phase markers, position-only plan map

Diagnosis (2026-09-11 user testing, minimal-effort GLM): sessions stuck on
'plan' forever. Two causes: (1) the three-phase view read the stored tree's
currentNodeId, which nothing ever writes post author-once (2026-09-03) — a
structural dead seam; (2) the teacher stochastically left quiz.nodeId
untagged, and done-by-quiz-result semantics would have falsely completed
nodes anyway.

The redesign (docs/probe-teach-split.md) abolishes derived occupancy in
favor of explicit events:

- Two agents over one transcript: createProberAgent (quiz without a
  nodeId field, ask, probe_done) and createTeacherAgent (full toolset,
  plan_init onward). The chat-handler construction branch keys on a
  probe_done part in the transcript. MODELS.prober added.
- probe_done: server-side seam tool — carries the learner assessment
  (strands + free-form edge read), refuses while a question is open or
  the probe is already concluded, and queues a handoff steer through the
  generalized wake pipeline (steers now dedupe by id, not jobId; job
  finalization only for job steers). The teacher auto-wakes and authors
  the plan; the learner sees the assessment via the probe-done card.
- phases: probe ends at a concluded probe_done, plan ends at a
  successful plan_init — strictly sequential, no skipping, no legacy
  compatibility (data wiped at the pivot per doctrine).
- Plan map = position, not completion: derivePlanStatuses reads only
  tagged quiz CALLS (state/outcome feed nothing); current = latest
  tagged node with consecutive duplicates collapsed, visited = earlier
  distinct stays (a revisit stays in both — current wins at render).
  PlanTree is structure-only; currentNodeId/doneNodeIds and the dead
  mutation helpers are gone. No fallback current.
- Prompts split per agent (prompts/prober/, prompts/teacher/ over the
  trimmed shared base); graph-context injection removed entirely
  (shelved, docs/probe-teach-split.md §5); knowledge-graph context no
  longer reaches either agent.
- Plan blocks are full-only snapshots at plan_init (neighborhood
  machinery retired; the persistent DAG window is the future UI wave);
  usage capture unchanged.

Gates: tsc clean, 294/294 tests (phases/plan-tree/plan-blocks/question-
tools rewritten to the new semantics; probe-done + handoff-wake tests
added), lint, build. Live GLM smoke PASS: prober concluded with an honest
assessment, teacher auto-woke, authored a 10-node plan with the block
attached, then tagged checks on slope-instinct → secant derived current/
visited exactly as designed; grading honest both ways (driver off-by-one
proved a wrong answer grades false).

components/artifacts/index.ts

@@ -3,6 +3,7 @@ import type { ArtifactPartProps } from "@/src/core";
 import { DemoCard } from "@/components/artifacts/demo-card";
 import { JobCard } from "@/components/artifacts/job-card";
 import { VisualCard } from "@/components/artifacts/visual-card";
+import { ProbeDoneCard } from "@/components/artifacts/probe-done-card";
 
 /**
  * Artifact renderer registry — the extension point for tool-part UI.
@@ -24,4 +25,7 @@ export const artifactRenderers: Record<string, ComponentType<ArtifactPartProps>>
 	// Artifacts (W6 fills in the components):
 	demo: DemoCard,
 	visual: VisualCard,
+	// The probe conclusion: the assessment belongs to the learner as much as
+	// to the transcript (docs/probe-teach-split.md §2). Plain v1 card.
+	probe_done: ProbeDoneCard,
 };

components/artifacts/probe-done-card.tsx

@@ -0,0 +1,48 @@
+"use client";
+
+import { ClipboardCheckIcon } from "lucide-react";
+import type { ProbeDoneResult, ProbeDoneInput } from "@/src/core";
+
+/**
+ * The probe conclusion card (docs/probe-teach-split.md §2): renders the
+ * prober's probe_done call. The INPUT is the learner assessment — strand by
+ * strand, where his edge sits — and it belongs to the learner as much as to
+ * the transcript: it streams live while the prober writes it and stays as
+ * the handoff artifact. The OUTPUT is the seam's ack (ok) or the refusal
+ * the prober retried on (rendered quiet and small — it is model-facing
+ * bookkeeping, kept only for transcript honesty).
+ *
+ * Deliberately plain v1: the proper assessment viewer design is deferred to
+ * the UI wave; this card just guarantees the data is never invisible.
+ */
+export function ProbeDoneCard({ part }: { part: { input?: unknown; output?: unknown } }) {
+	const input = part.input as ProbeDoneInput | undefined;
+	const output = part.output as ProbeDoneResult | undefined;
+	const assessment = output?.ok ? output.assessment : (input?.assessment ?? null);
+
+	return (
+		<div data-testid="probe-done-card" className="my-3 rounded-xl border bg-card/60 p-3">
+			<div className="mb-2 flex items-center gap-2 text-xs font-medium uppercase tracking-wide text-muted-foreground">
+				<ClipboardCheckIcon className="size-3.5 text-primary" />
+				Probe concluded — your assessment
+			</div>
+			{output && !output.ok ? (
+				<p className="text-sm text-muted-foreground">The prober retried: {output.error}</p>
+			) : assessment ? (
+				<div className="space-y-2">
+					{assessment.strands.map((s, i) => (
+						<div key={i} className="rounded-lg bg-background/60 px-3 py-2">
+							<div className="text-sm font-medium">{s.strand}</div>
+							<div className="mt-0.5 text-sm text-muted-foreground">{s.assessment}</div>
+						</div>
+					))}
+					{assessment.summary && (
+						<p className="px-1 text-sm text-muted-foreground">{assessment.summary}</p>
+					)}
+				</div>
+			) : (
+				<p className="text-sm text-muted-foreground">Writing the assessment…</p>
+			)}
+		</div>
+	);
+}