diff --git a/backend/planning_exercise_path_builder.py b/backend/planning_exercise_path_builder.py
index 384475a..5c2c700 100644
--- a/backend/planning_exercise_path_builder.py
+++ b/backend/planning_exercise_path_builder.py
@@ -1272,9 +1272,35 @@ def _match_roadmap_slot(
     else:
         step["slot_status"] = "matched"
         step["roadmap_match_source"] = "stage_spec"
+    if step.get("roadmap_match_source") != "slot_best_match" and not _roadmap_step_passes_post_match_gate(
+        cur,
+        step,
+        goal_query=goal_query,
+        semantic_brief=semantic_brief,
+    ):
+        return None, stage_spec
     return step, None
 
 
+def _roadmap_step_passes_post_match_gate(
+    cur,
+    step: Dict[str, Any],
+    *,
+    goal_query: str,
+    semantic_brief: PlanningSemanticBrief,
+) -> bool:
+    """Abgleich mit Pfad-QA — kein Rematch-Treffer, der sofort wieder stage_mismatch wäre."""
+    if step.get("exercise_id") is None:
+        return False
+    issues = detect_off_topic_steps(
+        cur,
+        [step],
+        brief=semantic_brief,
+        goal_query=goal_query,
+    )
+    return not issues
+
+
 def _normalize_roadmap_steps_coverage(
     steps: List[Dict[str, Any]],
     *,
@@ -1394,6 +1420,92 @@ def _purge_stage_mismatch_roadmap_slots(
     return out, new_unfilled
 
 
+def _enrich_roadmap_unfilled_gap_offers(
+    cur,
+    *,
+    steps: List[Dict[str, Any]],
+    gap_fill_offers: List[Dict[str, Any]],
+    body: ProgressionPathSuggestRequest,
+    roadmap_ctx: ProgressionRoadmapContext,
+    goal_query: str,
+    semantic_brief: PlanningSemanticBrief,
+) -> Tuple[List[Dict[str, Any]], List[Dict[str, Any]]]:
+    """KI-Lücken-Angebote für alle leeren Roadmap-Slots (nach Rematch/Normalize)."""
+    if not body.include_ai_gap_fill:
+        return steps, gap_fill_offers
+
+    seen_offer_ids = {o.get("offer_id") for o in gap_fill_offers if o.get("offer_id")}
+    out_steps: List[Dict[str, Any]] = []
+    offers = list(gap_fill_offers)
+
+    for raw in steps:
+        step = dict(raw)
+        if step.get("exercise_id") is not None:
+            out_steps.append(step)
+            continue
+        try:
+            major_idx = int(step["roadmap_major_step_index"])
+        except (TypeError, ValueError, KeyError):
+            out_steps.append(step)
+            continue
+        if step.get("gap_offer") and step.get("proposal_key"):
+            oid = step["gap_offer"].get("offer_id")
+            if oid and oid not in seen_offer_ids:
+                offers.append(dict(step["gap_offer"]))
+                seen_offer_ids.add(oid)
+            out_steps.append(step)
+            continue
+        stage_spec = next(
+            (
+                s
+                for s in (roadmap_ctx.stage_specs or [])
+                if int(s.major_step_index) == major_idx
+            ),
+            None,
+        )
+        learning_goal = (
+            (stage_spec.learning_goal if stage_spec else None)
+            or step.get("roadmap_learning_goal")
+            or step.get("title")
+            or ""
+        ).strip()
+        spec = {
+            "source": "roadmap_unfilled",
+            "insert_after_index": max(major_idx - 1, -1),
+            "roadmap_major_step_index": major_idx,
+            "phase": (step.get("roadmap_phase") or "vertiefung").strip().lower(),
+            "title_hint": (learning_goal or f"Slot {major_idx + 1}")[:120],
+            "sketch": learning_goal,
+            "rationale": (
+                f"Slot {major_idx + 1} — keine passende Bibliotheks-Übung; "
+                "KI-Entwurf für diese Stufe."
+            ),
+        }
+        offer = build_gap_fill_offer(
+            spec=spec,
+            steps=steps,
+            goal_query=goal_query,
+            brief=semantic_brief,
+            proposal=None,
+            roadmap_snapshot=_roadmap_gap_snapshot_for_spec(
+                cur,
+                roadmap_ctx,
+                spec,
+                goal_query=goal_query,
+                semantic_brief=semantic_brief,
+            ),
+        )
+        step["gap_offer"] = offer
+        step["proposal_key"] = offer.get("offer_id")
+        step["slot_status"] = "unfilled"
+        if offer.get("offer_id") and offer.get("offer_id") not in seen_offer_ids:
+            offers.append(offer)
+            seen_offer_ids.add(offer.get("offer_id"))
+        out_steps.append(step)
+
+    return out_steps, offers
+
+
 def _merge_rematch_unfilled(
     roadmap_unfilled: List[Tuple[int, StageSpecArtifact]],
     rematch_new_unfilled: List[Tuple[int, StageSpecArtifact]],
@@ -2269,7 +2381,7 @@ def suggest_progression_path(
         elif gaps and roadmap_first:
             unfilled_gaps = list(gaps)
 
-        if body.include_llm_path_qa:
+        if body.include_llm_path_qa and not roadmap_first:
             llm_qa, llm_qa_applied = try_llm_qa_progression_path(
                 cur,
                 goal_query=goal_query,
@@ -2348,6 +2460,22 @@ def suggest_progression_path(
                     roadmap_first=roadmap_first,
                 )
 
+        if body.include_llm_path_qa and roadmap_first:
+            gaps = detect_path_gaps(
+                cur,
+                steps,
+                brief=semantic_brief,
+                roadmap_first=roadmap_first,
+            )
+            llm_qa, llm_qa_applied = try_llm_qa_progression_path(
+                cur,
+                goal_query=goal_query,
+                brief=semantic_brief,
+                steps=steps,
+                gaps=gaps,
+                bridge_inserts=bridge_inserts,
+            )
+
         llm_gap_specs = parse_llm_suggested_new_exercises(
             llm_qa,
             brief=semantic_brief,
@@ -2397,6 +2525,22 @@ def suggest_progression_path(
             if offer.get("offer_id") not in seen_offer_ids:
                 gap_fill_offers.append(offer)
 
+    if roadmap_first and roadmap_ctx is not None:
+        steps = _normalize_roadmap_steps_coverage(
+            steps,
+            roadmap_ctx=roadmap_ctx,
+            max_steps=max_steps,
+        )
+        steps, gap_fill_offers = _enrich_roadmap_unfilled_gap_offers(
+            cur,
+            steps=steps,
+            gap_fill_offers=gap_fill_offers,
+            body=body,
+            roadmap_ctx=roadmap_ctx,
+            goal_query=goal_query,
+            semantic_brief=semantic_brief,
+        )
+
     multistage_qa = run_multistage_path_qa(
         off_topic_steps=off_topic_steps,
         stripped_off_topic=stripped_off_topic,
@@ -2428,71 +2572,6 @@ def suggest_progression_path(
         path_qa["refine_log"] = refine_log
         path_qa["refine_count"] = len(refine_log)
 
-    if roadmap_first and roadmap_ctx is not None:
-        steps = _normalize_roadmap_steps_coverage(
-            steps,
-            roadmap_ctx=roadmap_ctx,
-            max_steps=max_steps,
-        )
-        if body.include_ai_gap_fill:
-            seen_offer_ids = {o.get("offer_id") for o in gap_fill_offers if o.get("offer_id")}
-            for step in steps:
-                if step.get("exercise_id") is not None:
-                    continue
-                try:
-                    major_idx = int(step["roadmap_major_step_index"])
-                except (TypeError, ValueError, KeyError):
-                    continue
-                if step.get("gap_offer") and step.get("proposal_key"):
-                    oid = step["gap_offer"].get("offer_id")
-                    if oid and oid not in seen_offer_ids:
-                        gap_fill_offers.append(dict(step["gap_offer"]))
-                        seen_offer_ids.add(oid)
-                    continue
-                stage_spec = next(
-                    (
-                        s
-                        for s in (roadmap_ctx.stage_specs or [])
-                        if int(s.major_step_index) == major_idx
-                    ),
-                    None,
-                )
-                learning_goal = (
-                    (stage_spec.learning_goal if stage_spec else None)
-                    or step.get("roadmap_learning_goal")
-                    or step.get("title")
-                    or ""
-                ).strip()
-                spec = {
-                    "source": "roadmap_unfilled",
-                    "insert_after_index": max(major_idx - 1, -1),
-                    "roadmap_major_step_index": major_idx,
-                    "phase": (step.get("roadmap_phase") or "vertiefung").strip().lower(),
-                    "title_hint": (learning_goal or f"Slot {major_idx + 1}")[:120],
-                    "sketch": learning_goal,
-                    "rationale": f"Slot {major_idx + 1} — keine passende Bibliotheks-Übung; KI-Entwurf für diese Stufe.",
-                }
-                offer = build_gap_fill_offer(
-                    spec=spec,
-                    steps=steps,
-                    goal_query=goal_query,
-                    brief=semantic_brief,
-                    proposal=None,
-                    roadmap_snapshot=_roadmap_gap_snapshot_for_spec(
-                        cur,
-                        roadmap_ctx,
-                        spec,
-                        goal_query=goal_query,
-                        semantic_brief=semantic_brief,
-                    ),
-                )
-                step["gap_offer"] = offer
-                step["proposal_key"] = offer.get("offer_id")
-                step["slot_status"] = "unfilled"
-                if offer.get("offer_id") and offer.get("offer_id") not in seen_offer_ids:
-                    gap_fill_offers.append(offer)
-                    seen_offer_ids.add(offer.get("offer_id"))
-
     filled_library_steps = sum(1 for s in steps if s.get("exercise_id") is not None)
     match_summary = {
         "roadmap_first": roadmap_first,
diff --git a/backend/planning_path_rematch.py b/backend/planning_path_rematch.py
index dc042ee..200c3c7 100644
--- a/backend/planning_path_rematch.py
+++ b/backend/planning_path_rematch.py
@@ -207,8 +207,29 @@ def rematch_roadmap_slots(
                 }
             )
         else:
+            goal = (stage_spec.learning_goal or "").strip()
+            major = None
+            if roadmap_ctx.roadmap:
+                major = next(
+                    (m for m in roadmap_ctx.roadmap.major_steps if int(m.index) == int(major_idx)),
+                    None,
+                )
+            steps_by_major[int(major_idx)] = {
+                "exercise_id": None,
+                "variant_id": None,
+                "title": goal or f"Slot {major_idx + 1}",
+                "is_ai_proposal": False,
+                "roadmap_major_step_index": int(major_idx),
+                "roadmap_phase": major.phase if major else None,
+                "roadmap_learning_goal": goal or None,
+                "roadmap_match_source": "unfilled",
+                "slot_status": "unfilled",
+                "reasons": ["Keine passende Übung für Roadmap-Stufe"],
+            }
             if unfilled_spec is not None:
                 new_unfilled.append((step_index, unfilled_spec))
+            elif stage_spec is not None:
+                new_unfilled.append((step_index, stage_spec))
             rematch_log.append(
                 {
                     "roadmap_major_step_index": int(major_idx),
diff --git a/backend/tests/test_planning_path_rematch.py b/backend/tests/test_planning_path_rematch.py
index 78cc99f..aaeddb5 100644
--- a/backend/tests/test_planning_path_rematch.py
+++ b/backend/tests/test_planning_path_rematch.py
@@ -183,3 +183,43 @@ def test_rematch_excludes_replaced_exercise_from_used():
         match_slot_fn=_fake_match,
     )
     assert 99 in seen_used[0]
+
+
+def test_rematch_unfilled_leaves_placeholder_step():
+    specs = _stage_specs()
+    ctx = ProgressionRoadmapContext(
+        goal_query="Mae Geri",
+        max_steps=3,
+        stage_specs=specs,
+    )
+    steps = [
+        {"exercise_id": 10, "title": "OK", "roadmap_major_step_index": 0},
+        {"exercise_id": 99, "title": "Falsch", "roadmap_major_step_index": 1},
+    ]
+
+    def _no_match(cur, *, stage_spec, **kwargs):
+        return None, stage_spec
+
+    ordered, log, unfilled = rematch_roadmap_slots(
+        None,
+        tenant=None,
+        body=None,
+        goal_query="Mae Geri",
+        max_steps=3,
+        semantic_brief=None,
+        path_target_profile=None,
+        path_intent="",
+        roadmap_ctx=ctx,
+        steps=steps,
+        slot_indices={1},
+        rematch_reasons={1: "stage_mismatch"},
+        match_slot_fn=_no_match,
+    )
+
+    assert len(ordered) == 2
+    slot1 = ordered[1]
+    assert slot1["exercise_id"] is None
+    assert slot1["slot_status"] == "unfilled"
+    assert slot1["roadmap_match_source"] == "unfilled"
+    assert log[0]["action"] == "rematch_unfilled"
+    assert len(unfilled) == 1
diff --git a/backend/version.py b/backend/version.py
index 17b1603..ce251cc 100644
--- a/backend/version.py
+++ b/backend/version.py
@@ -1,6 +1,6 @@
 # Shinkan Jinkendo Version Information
 
-APP_VERSION = "0.8.230"
+APP_VERSION = "0.8.231"
 BUILD_DATE = "2026-05-22"
 DB_SCHEMA_VERSION = "20260607090"
 
@@ -38,7 +38,7 @@ MODULE_VERSIONS = {
     "skill_profiles": "1.0.0",  # Phase 3: gewichtetes Fähigkeiten-Profil + skill-discovery/suggestions
     "methods": "0.1.0",
     "exercises": "2.37.1",  # KI-Endpoints: feature_usage nach ai_calls consume
-    "planning_exercise_suggest": "0.23.5",  # Roadmap-Match strikt; stage_mismatch → unfilled + KI-Gap
+    "planning_exercise_suggest": "0.23.6",  # Gap-Angebote nach Rematch; LLM-QA auf finalem Pfad; Post-Match-Gate
     "training_units": "0.4.0",  # POST .../publish-to-framework: Ablauf aus geplanter Einheit → Rahmen-Slot-Blueprint
     "training_programs": "0.1.0",
     "planning": "0.15.0",  # Vorlagen: Strukturvorschau, Bearbeiten inkl. Split-Sessions + Beschreibung
diff --git a/frontend/src/utils/progressionGraphDraft.js b/frontend/src/utils/progressionGraphDraft.js
index 1a6489e..908cba8 100644
--- a/frontend/src/utils/progressionGraphDraft.js
+++ b/frontend/src/utils/progressionGraphDraft.js
@@ -856,21 +856,16 @@ export function slotsToEvaluateSteps(draft) {
 
 export function applyMatchStepsToSlots(draft, apiSteps) {
   const steps = Array.isArray(apiSteps) ? apiSteps : []
-  const nextSlots = (draft.slots || []).map((slot) => ({
-    ...slot,
-    primary: { ...slot.primary },
-    siblings: [...(slot.siblings || [])],
-  }))
-
-  const touchedMajors = new Set()
+  const stepByMajor = new Map()
   for (const step of steps) {
     if (step.roadmap_major_step_index == null || !Number.isFinite(Number(step.roadmap_major_step_index))) {
       continue
     }
-    const idx = Number(step.roadmap_major_step_index)
-    if (idx < 0 || idx >= nextSlots.length) continue
-    touchedMajors.add(idx)
+    stepByMajor.set(Number(step.roadmap_major_step_index), step)
+  }
 
+  const mapStepToPrimary = (step, slot) => {
+    const midx = Number(slot.majorStepIndex)
     const isProposal = Boolean(step.is_ai_proposal) || step.exercise_id == null
     const hasAiPayload = Boolean(step.ai_suggestion) || Boolean(step.proposal_key)
     const isUnfilledSlot =
@@ -880,33 +875,50 @@ export function applyMatchStepsToSlots(draft, apiSteps) {
       Boolean(step.gap_offer)
     if (isProposal && !hasAiPayload && isUnfilledSlot) {
       const offer = step.gap_offer || {}
-      nextSlots[idx].primary = proposalSlotExercise({
+      return proposalSlotExercise({
         title:
           offer.title_hint ||
           step.roadmap_learning_goal ||
           step.title ||
-          nextSlots[idx].learning_goal ||
-          `Slot ${idx + 1}`,
-        proposalKey: offer.offer_id || step.proposal_key || `roadmap-unfilled-${idx}`,
+          slot.learning_goal ||
+          `Slot ${midx + 1}`,
+        proposalKey: offer.offer_id || step.proposal_key || `roadmap-unfilled-${midx}`,
         aiSuggestion: offer.ai_suggestion || null,
       })
-    } else if (isProposal && !hasAiPayload) {
-      nextSlots[idx].primary = emptySlotExercise()
-    } else if (isProposal) {
-      nextSlots[idx].primary = proposalSlotExercise({
-        title: step.title || nextSlots[idx].learning_goal,
+    }
+    if (isProposal && !hasAiPayload) {
+      return emptySlotExercise()
+    }
+    if (isProposal) {
+      return proposalSlotExercise({
+        title: step.title || slot.learning_goal,
         proposalKey: step.proposal_key,
         aiSuggestion: step.ai_suggestion,
       })
-    } else {
-      nextSlots[idx].primary = librarySlotExercise({
-        exerciseId: step.exercise_id,
-        exerciseTitle: step.title || `Übung #${step.exercise_id}`,
-        variantId: step.variant_id,
-      })
     }
+    return librarySlotExercise({
+      exerciseId: step.exercise_id,
+      exerciseTitle: step.title || `Übung #${step.exercise_id}`,
+      variantId: step.variant_id,
+    })
   }
 
+  const nextSlots = (draft.slots || []).map((slot) => {
+    const base = {
+      ...slot,
+      primary: { ...slot.primary },
+      siblings: [...(slot.siblings || [])],
+    }
+    const step = stepByMajor.get(Number(slot.majorStepIndex))
+    if (!step) {
+      return base
+    }
+    return {
+      ...base,
+      primary: mapStepToPrimary(step, slot),
+    }
+  })
+
   return syncProgressionRoadmapFromSlots({ ...draft, slots: nextSlots, dirty: true })
 }