practice: pose-time target validation — names normalize to ids, unknowns rejected (round 9 slice 2)

problemTool takes the whole-graph node source (chat-handler passes
graphStore.readNodes, read fresh per pose): prepareArguments silently
normalizes name targets to canonical ids so the RECORD carries ids
(the mastery trace groups on them); execute rejects unresolvable
targets with a one-line invalidInputResult — no re-listing, the [id]
proposition list already sits in the agent's context. Validation set
is the whole graph, not the picks: pickless sessions and future
graph-tools poses stay valid. Absent node source (tests) = no
validation.

src/server/chat-handler.ts

@@ -328,7 +328,18 @@ export async function handleChatInput(
 	const constructed: TeachingAgent = buildAgent
 		? buildAgent({ ...agentOpts, teaching })
 		: record.kind === "practice"
-			? createPracticeAgent({ chatId, chatStore: stores.chatStore }, agentOpts)
+			? createPracticeAgent(
+					{
+						chatId,
+						chatStore: stores.chatStore,
+						// TARGET VALIDATION (owner round 9): the practice tools read the
+						// WHOLE graph fresh per pose — targets validate against every
+						// proposition, not just the session's picks (pickless sessions
+						// and future graph-tools poses stay valid).
+						graphNodes: () => stores.graphStore.readNodes(),
+					},
+					agentOpts,
+				)
 			: teaching
 			? createTeacherAgent(
 					{