update

templates/Prompt/Blocks-Gruppierung.md

@@ -1,6 +1,6 @@
 Topic "{topic}". A previous step produced a flat list of learning items that is TOO FINE-GRAINED for a table of contents — single environment variables, single CLI flags, single config fields, near-synonyms. Your job is **card sorting**: group the items into a small set of **theme blocks**, so that EVERY item lands in exactly one theme. Nothing is discarded — grouping never loses an item; it only organises them.
 
-Think of the result as chapters: ~15–25 themes for a whole topic, each a coherent learning block that a learner meets as one unit. A theme with only one fitting item is fine (a genuinely standalone concept), but prefer pulling related items together over leaving many singletons.
+Think of the result as chapters: ~{theme_lo}–{theme_hi} themes for a whole topic of this size ({n_items} items), each a coherent learning block that a learner meets as one unit. A theme with only one fitting item is fine (a genuinely standalone concept), but prefer pulling related items together over leaving many singletons.
 
 CANDIDATES you may group (your starting point — a pre-clustered neighbourhood):
 {candidates}

templates/Prompt/Blocks-Luecken-Research.md

@@ -0,0 +1,20 @@
+Topic "{topic}". A quality audit found GAPS: source-material passages and named results that NO existing inventory block covers. Your job: propose new learning blocks ONLY for these gaps — nothing else. An empty list is a valid answer (when the gaps are filler or already covered under another name).
+
+EXISTING BLOCKS (do NOT re-propose these or near-synonyms of them):
+{blocks}
+
+UNCOVERED MATERIAL (gap passages and named results, with source excerpts):
+{gaps}
+
+Rules:
+- One block per genuinely missing, teachable concept — self-contained, exam-relevant.
+- Ground every proposal in the material above; never invent canon knowledge the material does not treat.
+- Title: concrete concept name (max 8 words, no catalog references like "(Satz 6.33)").
+- Description: ONE sentence naming what the block teaches, grounded in the material.
+- Skip organizational text, prefaces, exercise scaffolding, duplicates of existing blocks.
+
+Write ONLY the JSON file to: {out_path}
+
+Format (`blocks` may be empty):
+{{"blocks": [{{"title": "…", "description": "…"}}]}}
+{extra}

templates/Prompt/Subblock-Crossblock.md

@@ -1,13 +0,0 @@
-Das Lernbaustein-Inventar zum Thema "{topic}" ist in Blöcke mit je eigenen Subbausteinen zerlegt. Manche Aussage taucht in ZWEI Blöcken auf — im Lernguide steht sie dann doppelt. Für jedes Paar unten: Treffen A und B DIESELBE Aussage?
-
-PAARE (jeweils mit Block-Zugehörigkeit und Kernpunkten):
-{pairs}
-
-## Entscheidung pro Paar
-- **DIESELBE Aussage** (auch anders formuliert, oder eine ist Teilmenge der anderen): Welcher Block ist die natürliche Heimat der Aussage? → antworte **"a"** (A behält sie, B verliert sie) oder **"b"** (B behält sie, A verliert sie). Heimat ist der Block, in dessen Kernthema die Aussage gehört — nicht der, der sie nur am Rand streift.
-- **VERSCHIEDENE Aussagen** → antworte **"nein"**. Dazu zählt: gleiche Regel, aber auf VERSCHIEDENE Kontexte angewendet (die Anwendung im jeweiligen Block-Kontext ist eigener Lernstoff); Grundregel vs. Sonderfall; gegensätzliche Aussagen.
-- Im Zweifel: **"nein"**.
-
-Antworte NUR mit JSON, ohne weiteren Text (jede Paar-Nummer mit "a", "b" oder "nein"):
-{{"pairs": {{"1": "a", "2": "nein"}}}}
-{extra}

templates/QA/QA-Konzept-Luecken.md

@@ -0,0 +1,15 @@
+You are an INDEPENDENT quality auditor for a learning-block inventory on the topic "{topic}". Below are NAMED results/concepts a scan found in the source material (named theorems, algorithms, definitions) that seem to map to NO inventory block. For each: is this a real, exam-relevant concept the inventory is missing → **ja**, or is it covered by an existing block under another name / a mangled-extraction artefact / mere organizational labeling → **nein**?
+
+NAMED RESULTS WITHOUT A COVERING BLOCK:
+{results}
+
+Rules:
+- Covered by an existing block under a synonym, symbol variant or broader umbrella → nein.
+- Garbled extraction artefacts (split or shredded words) and pure numbering references → nein.
+- A genuine named result the material teaches and no block captures → ja.
+- When genuinely unsure → ja (a missed gap is worse than one extra research round).
+
+Reply with ONLY the JSON — no other text, no code fences.
+Format (one verdict per number):
+{{"relevant": {{"1": "ja", "2": "nein"}}}}
+{extra}