feat(B-LLM): Zentraler LLM Client — llm_complete() + llm_embed() + Migration + Tests + Doku

B-LLM: llm_client.py um generische llm_complete() und llm_embed() erweitert
- Provider-Auswahl, API-Key-Auflösung, Error-Handling, Cost-Tracking
- Retry mit Exponential-Backoff für transient errors
- Timeout konfigurierbar
- Helper: get_api_credentials(), build_model(), _classify_error()

B-LLM-MIG: Alle 8 direkten litellm.acompletion() Calls auf llm_complete() umgestellt
- agent_runner.py, query_understanding.py (2x), ai_proactive (3x), ai_assistant (2x)
- 0 verbleibende direkte litellm.acompletion() Calls außerhalb llm_client.py

B-LLM-TEST: 39 Tests in test_llm_client.py — alle grün
- Mock mode, error handling, embed, helpers, backward compat

B-LLM-DOC: Plugin-Dev-Guide Kapitel 7 (LLM Integration) hinzugefügt

app/plugins/builtins/ai_proactive/jobs.py

@@ -16,7 +16,7 @@ import logging
 import uuid
 from typing import Any
 
-import litellm
+from app.ai.llm_client import llm_complete
 from sqlalchemy import select
 
 from app.core.db import create_db_session
@@ -201,29 +201,25 @@ async def deep_analysis(
             model_parts = model.split("/", 1)
             model = f"{provider_type}/{model_parts[-1]}"
 
-        litellm_kwargs: dict[str, Any] = dict(
-            model=model,
-            messages=[
-                {"role": "system", "content": DEEP_SYSTEM_PROMPT},
-                {
-                    "role": "user",
-                    "content": json.dumps(
-                        extended_context, default=str, ensure_ascii=False
-                    ),
-                },
-            ],
-            temperature=0.3,
-            max_tokens=1500,
-            response_format={"type": "json_object"},
-        )
-        if api_key:
-            litellm_kwargs["api_key"] = api_key
-        if api_base:
-            litellm_kwargs["api_base"] = api_base
-
         try:
-            response = await litellm.acompletion(**litellm_kwargs)
-            content = response.choices[0].message.content
+            result = await llm_complete(
+                model=model,
+                messages=[
+                    {"role": "system", "content": DEEP_SYSTEM_PROMPT},
+                    {
+                        "role": "user",
+                        "content": json.dumps(
+                            extended_context, default=str, ensure_ascii=False
+                        ),
+                    },
+                ],
+                temperature=0.3,
+                max_tokens=1500,
+                response_format={"type": "json_object"},
+                api_key=api_key,
+                api_base=api_base,
+            )
+            content = result["content"]
             if not content:
                 return
             # Strip markdown code fences if present