Release: rate limit key derivation fix + CORS security headers (#180)

Release: rate limit key derivation fix + CORS security headers

api/alembic/env.py

@@ -53,6 +53,7 @@ def run_migrations_online() -> None:
         # checkfirst=True ensures this is a no-op on existing databases.
         try:
             Base.metadata.create_all(bind=connection, checkfirst=True)
+            connection.commit()
         except Exception as exc:
             import logging
             logging.getLogger("alembic.env").warning(

api/alembic/versions/008_create_domain_tables.py

@@ -0,0 +1,210 @@
+"""Create domain tables (stores, purchases, coupons, etc.).
+
+Revision ID: 008_create_domain_tables
+Revises: 007_bootstrap_users_table
+Create Date: 2026-04-04
+"""
+
+import sqlalchemy as sa
+from sqlalchemy import text
+
+from alembic import op
+
+revision = "008_create_domain_tables"
+down_revision = "007_bootstrap_users_table"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+
+    # 1. stores
+    if not inspector.has_table("stores"):
+        op.create_table(
+            "stores",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("name", sa.String(100), nullable=False),
+            sa.Column("slug", sa.String(20), nullable=False, unique=True),
+            sa.Column("logo_url", sa.String(500), nullable=True),
+            sa.Column("website_url", sa.String(500), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+        )
+
+    # 2. store_locations
+    if not inspector.has_table("store_locations"):
+        op.create_table(
+            "store_locations",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("store_id", sa.Uuid(), sa.ForeignKey("stores.id"), nullable=False),
+            sa.Column("address", sa.String(300), nullable=False),
+            sa.Column("city", sa.String(100), nullable=False),
+            sa.Column("state", sa.String(2), nullable=False),
+            sa.Column("zip", sa.String(10), nullable=False),
+            sa.Column("lat", sa.Float(), nullable=True),
+            sa.Column("lng", sa.Float(), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+        )
+
+    # 3. normalized_products
+    if not inspector.has_table("normalized_products"):
+        op.create_table(
+            "normalized_products",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("canonical_name", sa.String(300), nullable=False),
+            sa.Column("category", sa.String(50), nullable=True),
+            sa.Column("subcategory", sa.String(100), nullable=True),
+            sa.Column("brand", sa.String(200), nullable=True),
+            sa.Column("size", sa.String(50), nullable=True),
+            sa.Column("size_unit", sa.String(10), nullable=True),
+            sa.Column("upc_variants", sa.JSON(), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+        )
+
+    # 4. purchases
+    if not inspector.has_table("purchases"):
+        op.create_table(
+            "purchases",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("user_id", sa.Text(), sa.ForeignKey("users.id"), nullable=False),
+            sa.Column("store_id", sa.Uuid(), sa.ForeignKey("stores.id"), nullable=False),
+            sa.Column("store_location_id", sa.Uuid(), sa.ForeignKey("store_locations.id"), nullable=True),
+            sa.Column("receipt_id", sa.String(200), nullable=False),
+            sa.Column("purchase_date", sa.Date(), nullable=False),
+            sa.Column("total", sa.Numeric(10, 2), nullable=False),
+            sa.Column("subtotal", sa.Numeric(10, 2), nullable=True),
+            sa.Column("tax", sa.Numeric(10, 2), nullable=True),
+            sa.Column("savings_total", sa.Numeric(10, 2), nullable=True),
+            sa.Column("source_url", sa.String(500), nullable=True),
+            sa.Column("raw_data", sa.JSON(), nullable=True),
+            sa.Column("ingested_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.UniqueConstraint("user_id", "store_id", "receipt_id", name="uq_purchase_receipt"),
+            sa.Index("ix_purchases_user_store", "user_id", "store_id"),
+        )
+
+    # 5. purchase_items
+    if not inspector.has_table("purchase_items"):
+        op.create_table(
+            "purchase_items",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("purchase_id", sa.Uuid(), sa.ForeignKey("purchases.id"), nullable=False),
+            sa.Column("product_name_raw", sa.String(300), nullable=False),
+            sa.Column("upc", sa.String(20), nullable=True),
+            sa.Column("quantity", sa.Numeric(10, 3), nullable=False),
+            sa.Column("unit_price", sa.Numeric(10, 2), nullable=False),
+            sa.Column("extended_price", sa.Numeric(10, 2), nullable=False),
+            sa.Column("regular_price", sa.Numeric(10, 2), nullable=True),
+            sa.Column("sale_price", sa.Numeric(10, 2), nullable=True),
+            sa.Column("coupon_discount", sa.Numeric(10, 2), nullable=True),
+            sa.Column("loyalty_discount", sa.Numeric(10, 2), nullable=True),
+            sa.Column("category_raw", sa.String(100), nullable=True),
+            sa.Column("normalized_product_id", sa.Uuid(), sa.ForeignKey("normalized_products.id"), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+        )
+
+    # 6. coupons
+    if not inspector.has_table("coupons"):
+        op.create_table(
+            "coupons",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("store_id", sa.Uuid(), sa.ForeignKey("stores.id"), nullable=False),
+            sa.Column("normalized_product_id", sa.Uuid(), sa.ForeignKey("normalized_products.id"), nullable=True),
+            sa.Column("title", sa.String(300), nullable=False),
+            sa.Column("description", sa.String(1000), nullable=True),
+            sa.Column("discount_type", sa.String(20), nullable=False),
+            sa.Column("discount_value", sa.Numeric(10, 2), nullable=True),
+            sa.Column("min_purchase", sa.Numeric(10, 2), nullable=True),
+            sa.Column("valid_from", sa.Date(), nullable=True),
+            sa.Column("valid_to", sa.Date(), nullable=True),
+            sa.Column("requires_clip", sa.Boolean(), server_default=text("false"), nullable=False),
+            sa.Column("coupon_code", sa.String(100), nullable=True),
+            sa.Column("source_url", sa.String(500), nullable=True),
+            sa.Column("scraped_at", sa.DateTime(timezone=True), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+        )
+
+    # 7. price_history
+    if not inspector.has_table("price_history"):
+        op.create_table(
+            "price_history",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("normalized_product_id", sa.Uuid(), sa.ForeignKey("normalized_products.id"), nullable=False),
+            sa.Column("store_id", sa.Uuid(), sa.ForeignKey("stores.id"), nullable=False),
+            sa.Column("observed_date", sa.Date(), nullable=False),
+            sa.Column("regular_price", sa.Numeric(10, 2), nullable=False),
+            sa.Column("sale_price", sa.Numeric(10, 2), nullable=True),
+            sa.Column("loyalty_price", sa.Numeric(10, 2), nullable=True),
+            sa.Column("coupon_price", sa.Numeric(10, 2), nullable=True),
+            sa.Column("source", sa.String(20), nullable=False),
+            sa.Column("purchase_item_id", sa.Uuid(), sa.ForeignKey("purchase_items.id"), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Index("ix_price_history_product_store_date", "normalized_product_id", "store_id", "observed_date"),
+        )
+
+    # 8. shrinkflation_events
+    if not inspector.has_table("shrinkflation_events"):
+        op.create_table(
+            "shrinkflation_events",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("normalized_product_id", sa.Uuid(), sa.ForeignKey("normalized_products.id"), nullable=False),
+            sa.Column("detected_date", sa.Date(), nullable=False),
+            sa.Column("old_size", sa.String(50), nullable=False),
+            sa.Column("new_size", sa.String(50), nullable=False),
+            sa.Column("old_unit", sa.String(10), nullable=True),
+            sa.Column("new_unit", sa.String(10), nullable=True),
+            sa.Column("price_at_old_size", sa.Numeric(10, 2), nullable=True),
+            sa.Column("price_at_new_size", sa.Numeric(10, 2), nullable=True),
+            sa.Column("confidence", sa.Numeric(3, 2), server_default=text("1.00"), nullable=False),
+            sa.Column("notes", sa.String(1000), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+        )
+
+    # 9. user_store_accounts
+    if not inspector.has_table("user_store_accounts"):
+        op.create_table(
+            "user_store_accounts",
+            sa.Column("id", sa.Uuid(), server_default=text("gen_random_uuid()"), primary_key=True),
+            sa.Column("user_id", sa.Text(), sa.ForeignKey("users.id"), nullable=False),
+            sa.Column("store_id", sa.Uuid(), sa.ForeignKey("stores.id"), nullable=False),
+            sa.Column("session_data", sa.JSON(), nullable=True),
+            sa.Column("session_expires_at", sa.DateTime(timezone=True), nullable=True),
+            sa.Column("last_sync_at", sa.DateTime(timezone=True), nullable=True),
+            sa.Column("status", sa.String(20), server_default=text("'active'"), nullable=False),
+            sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), nullable=False),
+            sa.UniqueConstraint("user_id", "store_id", name="uq_user_store_account"),
+        )
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+
+    if inspector.has_table("user_store_accounts"):
+        op.drop_table("user_store_accounts")
+    if inspector.has_table("shrinkflation_events"):
+        op.drop_table("shrinkflation_events")
+    if inspector.has_table("price_history"):
+        op.drop_table("price_history")
+    if inspector.has_table("coupons"):
+        op.drop_table("coupons")
+    if inspector.has_table("purchase_items"):
+        op.drop_table("purchase_items")
+    if inspector.has_table("purchases"):
+        op.drop_table("purchases")
+    if inspector.has_table("normalized_products"):
+        op.drop_table("normalized_products")
+    if inspector.has_table("store_locations"):
+        op.drop_table("store_locations")
+    if inspector.has_table("stores"):
+        op.drop_table("stores")

api/src/cartsnitch_api/config.py

@@ -13,14 +13,13 @@ class Settings(BaseSettings):
     )
     redis_url: str = "redis://localhost:6379/0"
 
-    jwt_secret_key: str = "change-me-in-production"
+    jwt_secret_key: str
     jwt_algorithm: str = "HS256"
     jwt_access_token_expire_minutes: int = 15
     jwt_refresh_token_expire_days: int = 7
 
-    service_key: str = "change-me-in-production"
-    # Valid Fernet key for local dev — MUST be overridden in production
-    fernet_key: str = "7reF42nmTwbdN21PBoubGp7h_FU8qSimstmlaMLoRK8="
+    service_key: str
+    fernet_key: str
 
     auth_service_url: str = "http://auth:3001"
 
@@ -35,9 +34,26 @@ class Settings(BaseSettings):
     rate_limit_window_seconds: int = 60
     rate_limit_enabled: bool = True
 
+    _PLACEHOLDER_VALUES = {"change-me-in-production"}
+
     @model_validator(mode="after")
-    def validate_fernet_key(self):
-        """Validate fernet_key is a valid 32-byte url-safe base64 key at startup."""
+    def validate_secrets(self):
+        if not self.jwt_secret_key or self.jwt_secret_key in self._PLACEHOLDER_VALUES:
+            raise ValueError(
+                "CARTSNITCH_JWT_SECRET_KEY must be set to a secure value. "
+                'Generate one with: python -c "import secrets; print(secrets.token_urlsafe(32))"'
+            )
+        if not self.service_key or self.service_key in self._PLACEHOLDER_VALUES:
+            raise ValueError(
+                "CARTSNITCH_SERVICE_KEY must be set to a secure value. "
+                'Generate one with: python -c "import secrets; print(secrets.token_urlsafe(32))"'
+            )
+        if not self.fernet_key or self.fernet_key in self._PLACEHOLDER_VALUES:
+            raise ValueError(
+                "CARTSNITCH_FERNET_KEY must be set to a valid Fernet key. "
+                "Generate one with: python -c "
+                "'from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())'"
+            )
         try:
             decoded = base64.urlsafe_b64decode(self.fernet_key.encode())
             if len(decoded) != 32:

api/src/cartsnitch_api/middleware/cors.py

@@ -11,6 +11,6 @@ def add_cors_middleware(app: FastAPI) -> None:
         CORSMiddleware,
         allow_origins=settings.cors_origins,
         allow_credentials=True,
-        allow_methods=["*"],
-        allow_headers=["*"],
+        allow_methods=["GET", "POST", "PUT", "DELETE", "PATCH", "OPTIONS"],
+        allow_headers=["Content-Type", "Authorization", "Accept", "Origin", "X-Requested-With"],
     )

api/src/cartsnitch_api/middleware/rate_limit.py

@@ -4,6 +4,7 @@ Uses in-memory sliding window as fallback, Redis/DragonflyDB when available.
 Per-IP limiting on public endpoints, per-token limiting on authenticated endpoints.
 """
 
+import hashlib
 import time
 from collections import defaultdict
 from threading import Lock
@@ -71,8 +72,8 @@ def _get_rate_limit_key(request: Request) -> tuple[str, _SlidingWindowCounter]:
     auth_header = request.headers.get("authorization", "")
     if auth_header.startswith("Bearer "):
         token = auth_header[7:]
-        # Use last 16 chars of token as key to avoid storing full tokens
-        return f"token:{token[-16:]}", _auth_limiter
+        token_hash = hashlib.sha256(token.encode()).hexdigest()
+        return f"token:{token_hash}", _auth_limiter
 
     # Fallback to IP for unauthenticated non-public endpoints
     return f"ip:{_get_client_ip(request)}", _public_limiter

api/tests/conftest.py

@@ -19,6 +19,25 @@ from cartsnitch_api.database import get_db
 from cartsnitch_api.main import create_app
 from cartsnitch_api.models import Base
 
+TEST_JWT_SECRET = secrets.token_urlsafe(32)
+TEST_SERVICE_KEY = secrets.token_urlsafe(32)
+TEST_FERNET_KEY = "7reF42nmTwbdN21PBoubGp7h_FU8qSimstmlaMLoRK8="
+
+
+@pytest.fixture(autouse=True)
+def setup_test_settings():
+    original_jwt = cartsnitch_settings.jwt_secret_key
+    original_service = cartsnitch_settings.service_key
+    original_fernet = cartsnitch_settings.fernet_key
+    cartsnitch_settings.jwt_secret_key = TEST_JWT_SECRET
+    cartsnitch_settings.service_key = TEST_SERVICE_KEY
+    cartsnitch_settings.fernet_key = TEST_FERNET_KEY
+    yield
+    cartsnitch_settings.jwt_secret_key = original_jwt
+    cartsnitch_settings.service_key = original_service
+    cartsnitch_settings.fernet_key = original_fernet
+
+
 TEST_DATABASE_URL = "sqlite+aiosqlite:///:memory:"
 
 
@@ -60,7 +79,8 @@ async def db_engine():
     async with engine.begin() as conn:
         await conn.run_sync(Base.metadata.create_all)
         # Create Better-Auth tables (not managed by SQLAlchemy models)
-        await conn.execute(text("""
+        await conn.execute(
+            text("""
             CREATE TABLE IF NOT EXISTS sessions (
                 id TEXT PRIMARY KEY,
                 token TEXT NOT NULL UNIQUE,
@@ -71,8 +91,10 @@ async def db_engine():
                 created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP NOT NULL,
                 updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP NOT NULL
             )
-        """))
-        await conn.execute(text("""
+        """)
+        )
+        await conn.execute(
+            text("""
             CREATE TABLE IF NOT EXISTS accounts (
                 id TEXT PRIMARY KEY,
                 user_id TEXT NOT NULL,
@@ -88,8 +110,10 @@ async def db_engine():
                 created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP NOT NULL,
                 updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP NOT NULL
             )
-        """))
-        await conn.execute(text("""
+        """)
+        )
+        await conn.execute(
+            text("""
             CREATE TABLE IF NOT EXISTS verifications (
                 id TEXT PRIMARY KEY,
                 identifier TEXT NOT NULL,
@@ -98,7 +122,8 @@ async def db_engine():
                 created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP NOT NULL,
                 updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP NOT NULL
             )
-        """))
+        """)
+        )
 
     yield engine
 
@@ -133,7 +158,9 @@ async def client(db_engine):
     app.dependency_overrides.clear()
 
 
-async def _create_test_user_and_session(client: AsyncClient, db_engine, **user_overrides) -> tuple[dict, str]:
+async def _create_test_user_and_session(
+    client: AsyncClient, db_engine, **user_overrides
+) -> tuple[dict, str]:
     """Create a test user and a valid session directly in the DB.
 
     Returns (user_dict, session_token).  Better-Auth stores the raw token

api/tests/test_middleware/test_rate_limit.py

@@ -1,8 +1,10 @@
 """Tests for rate limiting middleware."""
 
+from unittest.mock import MagicMock
+
 import pytest
 
-from cartsnitch_api.middleware.rate_limit import _SlidingWindowCounter
+from cartsnitch_api.middleware.rate_limit import _SlidingWindowCounter, _get_rate_limit_key
 
 
 class TestSlidingWindowCounter:
@@ -53,3 +55,32 @@ async def test_health_skips_rate_limit(client):
     resp = await client.get("/health")
     assert resp.status_code == 200
     assert "x-ratelimit-limit" not in resp.headers
+
+
+class TestGetRateLimitKey:
+    def _make_request(self, auth_header: str = "") -> MagicMock:
+        req = MagicMock()
+        req.url.path = "/purchases"
+        req.headers = {"authorization": auth_header} if auth_header else {}
+        return req
+
+    def test_distinct_tokens_produce_distinct_keys(self):
+        req1 = self._make_request("Bearer token_alpha_12345")
+        req2 = self._make_request("Bearer token_beta_67890")
+        key1, _ = _get_rate_limit_key(req1)
+        key2, _ = _get_rate_limit_key(req2)
+        assert key1 != key2
+
+    def test_same_token_produces_same_key(self):
+        req1 = self._make_request("Bearer same_token_value_abc")
+        req2 = self._make_request("Bearer same_token_value_abc")
+        key1, _ = _get_rate_limit_key(req1)
+        key2, _ = _get_rate_limit_key(req2)
+        assert key1 == key2
+
+    def test_key_does_not_contain_raw_token_suffix(self):
+        raw_token = "my_secret_jwt_token_xyz"
+        req = self._make_request(f"Bearer {raw_token}")
+        key, _ = _get_rate_limit_key(req)
+        assert raw_token[-16:] not in key
+        assert raw_token not in key

nginx.conf

@@ -9,6 +9,12 @@ server {
     gzip_types text/plain text/css application/json application/javascript text/xml application/xml application/xml+rss text/javascript image/svg+xml;
     gzip_min_length 256;
 
+    # Security headers
+    add_header X-Frame-Options "SAMEORIGIN" always;
+    add_header X-Content-Type-Options "nosniff" always;
+    add_header Referrer-Policy "strict-origin-when-cross-origin" always;
+    add_header Content-Security-Policy "default-src 'self'; script-src 'self'; style-src 'self' 'unsafe-inline'; img-src 'self' data: https:; font-src 'self'; connect-src 'self' https://*.cartsnitch.com https://*.farh.net; frame-ancestors 'self'" always;
+
     # Health endpoint for K8s probes
     location /health {
         access_log off;