Add LLM Council mode for multi-model consensus

3-stage council orchestration: parallel model queries (Stage 1),
anonymous peer ranking (Stage 2), and streamed chairman synthesis
(Stage 3). Includes scope-aware file resolution for Google/Claude
providers so upstream file attachments are visible to all providers.

- Backend: council.py orchestrator, /api/run_council_stream endpoint,
  query_model_full() non-streaming wrapper, resolve_provider() helper,
  resolve_scoped_file_ids() for Google/Claude scope parity with OpenAI
- Frontend: council toggle UI, model checkbox selector, chairman picker,
  SSE event parsing, tabbed Stage 1/2/3 response display
- Canvas: amber council node indicator with Users icon

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

1 files changed, 28 insertions, 1 deletions

diff --git a/backend/app/services/llm.py b/backend/app/services/llm.py
index 7efdce0..c22ada3 100644
--- a/backend/app/services/llm.py
+++ b/backend/app/services/llm.py
@@ -4,7 +4,7 @@ from typing import AsyncGenerator, List, Dict, Any, Optional
 import openai
 import google.generativeai as genai
 import anthropic
-from app.schemas import LLMConfig, Message, Role, Context
+from app.schemas import LLMConfig, Message, Role, Context, ModelProvider
 
 logger = logging.getLogger("contextflow.llm")
 
@@ -599,6 +599,33 @@ async def llm_streamer(
             yield f"Error calling LLM: {primary_error} (OpenRouter fallback also failed: {fallback_error})"
 
 
+def resolve_provider(model_name: str) -> ModelProvider:
+    """Determine the provider from a model name string."""
+    name = model_name.lower()
+    if any(name.startswith(p) for p in ('claude',)):
+        return ModelProvider.CLAUDE
+    if any(name.startswith(p) for p in ('gemini',)):
+        return ModelProvider.GOOGLE
+    # Default to OpenAI for gpt-*, o1, o3, etc.
+    return ModelProvider.OPENAI
+
+
+async def query_model_full(
+    context: Context,
+    user_prompt: str,
+    config: LLMConfig,
+    attachments=None,
+    tools=None,
+    openrouter_api_key=None,
+    images=None,
+) -> str:
+    """Collect full response from llm_streamer (non-streaming wrapper)."""
+    chunks = []
+    async for chunk in llm_streamer(context, user_prompt, config, attachments, tools, openrouter_api_key, images):
+        chunks.append(chunk)
+    return "".join(chunks)
+
+
 async def generate_title(user_prompt: str, response: str, api_key: str = None) -> str:
     """
     Generate a short title (3-4 words) for a Q-A pair using gpt-5-nano.