Add Knowledge Hub books FTS and remove training UI (0.16.0).

Index Assistent/books search.jsonl, expose knowledge hops/catalog in chat, and drop QLoRA/HF training stack.

Co-authored-by: Cursor <cursoragent@cursor.com>

Config/_base/patch-keys.json

@@ -8,6 +8,6 @@
     "snapshot_generate", "select_slot", "aspect", "images", "batch", "vary", "lock_seed",
     "creativity", "intensity", "complexity", "movement",
     "clear_prompt_images", "slot_to_prompt_image", "pack", "persona", "controls",
-    "inventory_query", "variants"
+    "inventory_query", "knowledge_query", "example_query", "knowledge_rating", "example_rating", "variants"
   ]
 }

Config/_base/skills/knowledge.json

@@ -0,0 +1,6 @@
+{
+  "id": "knowledge",
+  "title": "Books + knowledge search",
+  "default": false,
+  "prompt_file": "knowledge.md"
+}

Config/_base/skills/knowledge.md

@@ -0,0 +1,22 @@
+# Knowledge skill
+
+Attached **books** are read-only FTS corpora (Civitai prompts, RU fic style, …). They are **not** mutable memory.
+
+## When to search
+
+- User asks for reference prompts, scene wording, style/tone, or «как на Civitai».
+- You need concrete tag/prompt patterns before generating.
+
+## How to search (server hop)
+
+```json
+{ "ask": ["knowledge"], "knowledge_query": "redhead stockings cinematic window light" }
+```
+
+Optional: `knowledge_rating` (`pg`, `pg13`, …). Legacy `ask:["examples"]` + `example_query` still works for Civitai-only.
+
+## Rules
+
+- **Remix** — never paste long verbatim from books.
+- Pure chat / opinions: **prose only**, no JSON.
+- Do **not** emit `generate:true` on knowledge Q&A turns.

Config/_base/skills/memory.md

@@ -6,7 +6,7 @@ You have five memory tools:
 2. **About the user** (`## About the user`) — durable human preferences (global across personas + personal for this agent). Tunable weight in settings. Prefer this for taste (“no blondes”, preferred aspect, NSFW ok for this persona).
 3. **Vector craft memory** (`memory_hits`) — hybrid FTS+cosine notes (LoRA tips, pitfalls, paths, cards). Shared + this persona; personal overwrites shared on the same `kind`+`key`.
 4. **Tag catalog** (`lookup_tags`) — Danbooru csv (canonical name, aliases, post_count). **Not** RAG. Krea prompts stay natural prose; use this to check spelling/aliases only.
-5. **Civitai examples** (`ask: examples` / `lookup_examples`) — popular Krea 2 prompts indexed with FTS (no embeddings). Use when you want a **reference** for how others phrased a scene — remix, do not copy 1:1.
+5. **Civitai / knowledge books** (`ask: knowledge` / `ask: examples`) — FTS over attached books (gpu-rent seeded) and legacy Civitai examples. Remix, do not copy 1:1.
 
 ## Priority
 

Config/_base/training-qlora.json

@@ -1,49 +0,0 @@
-{
-  "hf_models": [
-    {
-      "id": "qwen2.5-7b-instruct",
-      "title": "Qwen2.5 7B Instruct",
-      "hf_id": "Qwen/Qwen2.5-7B-Instruct",
-      "ollama_hint": "qwen2.5:7b-instruct",
-      "default_output": "assistent-qwen25-7b:v1",
-      "rank": 16,
-      "seq_len": 2048
-    },
-    {
-      "id": "qwen2.5-3b-instruct",
-      "title": "Qwen2.5 3B Instruct",
-      "hf_id": "Qwen/Qwen2.5-3B-Instruct",
-      "ollama_hint": "qwen2.5:3b-instruct",
-      "default_output": "assistent-qwen25-3b:v1",
-      "rank": 16,
-      "seq_len": 2048
-    },
-    {
-      "id": "llama-3.2-3b-instruct",
-      "title": "Llama 3.2 3B Instruct",
-      "hf_id": "meta-llama/Llama-3.2-3B-Instruct",
-      "ollama_hint": "llama3.2:3b-instruct",
-      "default_output": "assistent-llama32-3b:v1",
-      "rank": 16,
-      "seq_len": 2048
-    },
-    {
-      "id": "llama-3.1-8b-instruct",
-      "title": "Llama 3.1 8B Instruct",
-      "hf_id": "meta-llama/Llama-3.1-8B-Instruct",
-      "ollama_hint": "llama3.1:8b-instruct",
-      "default_output": "assistent-llama31-8b:v1",
-      "rank": 16,
-      "seq_len": 2048
-    },
-    {
-      "id": "phi-3-mini-instruct",
-      "title": "Phi-3 Mini 4K Instruct",
-      "hf_id": "microsoft/Phi-3-mini-4k-instruct",
-      "ollama_hint": "phi3:mini",
-      "default_output": "assistent-phi3-mini:v1",
-      "rank": 16,
-      "seq_len": 2048
-    }
-  ]
-}