feat: complete phase 1 -- core framework with chat loop, agents, and React UI

Backend:
- FastAPI WebSocket /ws endpoint with streaming via LangGraph astream
- LangGraph Supervisor connecting 3 mock agents (order_lookup, order_actions, fallback)
- YAML Agent Registry with Pydantic validation and immutable configs
- PostgresSaver checkpoint persistence via langgraph-checkpoint-postgres
- Session TTL with 30-min sliding window and interrupt extension
- LLM provider abstraction (Anthropic/OpenAI/Google)
- Token usage + cost tracking callback handler
- Input validation: message size cap, thread_id format, content length
- Security: no hardcoded defaults, startup API key validation, no input reflection

Frontend:
- React 19 + TypeScript + Vite chat UI
- WebSocket hook with reconnect + exponential backoff
- Streaming token display with agent attribution
- Interrupt approval/reject UI for write operations
- Collapsible tool call viewer

Testing:
- 87 unit tests, 87% coverage (exceeds 80% requirement)
- Ruff lint + format clean

Infrastructure:
- Docker Compose (PostgreSQL 16 + backend)
- pyproject.toml with full dependency management

backend/app/config.py

@@ -0,0 +1,46 @@
+"""Centralized application configuration via pydantic-settings."""
+
+from __future__ import annotations
+
+from typing import Literal
+
+from pydantic import model_validator
+from pydantic_settings import BaseSettings, SettingsConfigDict
+
+
+class Settings(BaseSettings):
+    model_config = SettingsConfigDict(
+        env_file=".env",
+        env_file_encoding="utf-8",
+        extra="ignore",
+    )
+
+    database_url: str
+
+    llm_provider: Literal["anthropic", "openai", "google"] = "anthropic"
+    llm_model: str = "claude-sonnet-4-6"
+
+    session_ttl_minutes: int = 30
+    interrupt_ttl_minutes: int = 30
+
+    ws_host: str = "0.0.0.0"
+    ws_port: int = 8000
+
+    anthropic_api_key: str = ""
+    openai_api_key: str = ""
+    google_api_key: str = ""
+
+    @model_validator(mode="after")
+    def validate_provider_key(self) -> Settings:
+        key_map = {
+            "anthropic": self.anthropic_api_key,
+            "openai": self.openai_api_key,
+            "google": self.google_api_key,
+        }
+        key = key_map.get(self.llm_provider, "")
+        if not key:
+            raise ValueError(
+                f"API key for provider '{self.llm_provider}' is required. "
+                f"Set the corresponding environment variable."
+            )
+        return self