tradein-mvp/DEPLOY.md

@@ -57,7 +57,7 @@ import /opt/gendesign/tradein-mvp/deploy/Caddyfile.tradein-fragment
    shell-скриптом deploy через `source .env.runtime` перед `compose up`.
 2. `/opt/gendesign/tradein-mvp/backend/.env.runtime` — переменные внутри
    контейнера `tradein-backend` (читаются через `env_file:` в compose). Сюда
-   попадают `YANDEX_GEOCODER_API_KEY`, `ADMIN_TOKEN`, `COOKIE_ENCRYPTION_KEY` —
+   попадают `YANDEX_GEOCODER_API_KEY`, `COOKIE_ENCRYPTION_KEY` —
    всё, что нужно scripts/backfill_house_coords.py и application code внутри
    контейнера.
 
@@ -81,7 +81,6 @@ COOKIE_ENCRYPTION_KEY=<64-char hex>
 # Может быть симлинком на ../.env.runtime если переменные совпадают:
 #   ln -s ../.env.runtime /opt/gendesign/tradein-mvp/backend/.env.runtime
 YANDEX_GEOCODER_API_KEY=<key или пусто>
-ADMIN_TOKEN=<token или пусто>
 COOKIE_ENCRYPTION_KEY=<64-char hex>
 GENDESIGN_FDW_PASSWORD=<password или пусто>
 GLITCHTIP_DSN=<dsn или пусто>

tradein-mvp/backend/app/api/v1/admin.py

@@ -9,6 +9,7 @@ import json
 import logging
 import time
 from typing import Annotated, Any, Literal
+from urllib.parse import urlparse
 
 from fastapi import APIRouter, BackgroundTasks, Depends, HTTPException, Query
 from pydantic import BaseModel, Field
@@ -46,6 +47,17 @@ logger = logging.getLogger(__name__)
 router = APIRouter()
 
 
+def _assert_allowed_url(url: str) -> None:
+    """SSRF guard: отклоняем абсолютные URL с хостом не из allowlist (#756).
+
+    Относительные пути (urlparse().netloc пустой) пропускаем — хост
+    будет подставлен фиксированным AVITO_BASE / BASE_URL в самом scraper'е.
+    """
+    parsed = urlparse(url)
+    if parsed.netloc and parsed.netloc not in settings.scrape_allowed_hosts:
+        raise HTTPException(status_code=400, detail="host not allowed")
+
+
 _ALL_SOURCES = ["avito", "cian", "yandex", "n1"]
 
 
@@ -453,6 +465,7 @@ async def scrape_avito_house(
     Flow: fetch_house_catalog → save_house_catalog_enrichment.
     Returns counters {'house_id', 'reviews', 'sellers', 'listings_linked', 'placement_history'}.
     """
+    _assert_allowed_url(house_url)
     try:
         enrichment = await fetch_house_catalog(house_url)
     except Exception as e:
@@ -481,6 +494,7 @@ async def scrape_avito_detail(
     Flow: fetch_detail → save_detail_enrichment (UPDATE listings WHERE source='avito').
     Returns {'ok', 'item_id', 'updated'}.
     """
+    _assert_allowed_url(item_url)
     try:
         enrichment = await fetch_detail(item_url)
     except Exception as e:
@@ -929,6 +943,7 @@ async def scrape_yandex_detail(
     Returns a snapshot of the extracted DetailEnrichment (no DB write - read-only
     debug; main pipeline writes via estimator on /estimate flow).
     """
+    _assert_allowed_url(offer_url)
     async with YandexDetailScraper() as scraper:
         result = await scraper.fetch_detail(offer_url)
     if result is None:
@@ -1047,6 +1062,7 @@ async def scrape_cian_detail(
     If `listing_id` provided → save enrichment (offer_price_history + listings updates).
     Without it → debug-only (no DB write).
     """
+    _assert_allowed_url(offer_url)
     from app.services.scrapers.cian_detail import fetch_detail, save_detail_enrichment
 
     enrichment = await fetch_detail(offer_url)
@@ -1090,6 +1106,7 @@ async def scrape_cian_newbuilding(
     If `house_id` provided → persist (houses + houses_price_dynamics + management_companies).
     Without it → debug-only (no DB write).
     """
+    _assert_allowed_url(zhk_url)
     from app.services.scrapers.cian_newbuilding import (
         fetch_newbuilding,
         save_newbuilding_enrichment,

tradein-mvp/backend/app/core/config.py

@@ -108,5 +108,21 @@ class Settings(BaseSettings):
     # Конфигурируется через env ESTIMATE_QUOTA_LIMIT. Default 15.
     estimate_quota_limit: int = 15
 
+    # SSRF-защита для admin scrape endpoints (#756).
+    # Список хостов которым разрешено передавать абсолютные URL в параметрах *_url.
+    # Относительные пути (без netloc) проходят без проверки — хост подставляется
+    # фиксированным в самом scraper'е. Задаётся через env SCRAPE_ALLOWED_HOSTS
+    # (comma-separated). По умолчанию — все хосты используемые scrapers/*.py.
+    scrape_allowed_hosts: set[str] = {
+        "www.avito.ru",
+        "avito.ru",
+        "www.cian.ru",
+        "cian.ru",
+        "ekb.cian.ru",
+        "realty.yandex.ru",
+        "ekaterinburg.n1.ru",
+        "n1.ru",
+    }
+
 
 settings = Settings()

tradein-mvp/backend/tests/test_admin_ssrf.py

@@ -0,0 +1,106 @@
+"""Unit tests for SSRF allowlist guard (_assert_allowed_url) — issue #756.
+
+Approach: unit-tests on the helper directly (no DB/scraper fixtures needed).
+The guard is a pure synchronous function; TestClient integration is not required
+because the guard raises before any async scraper is called.
+"""
+
+from __future__ import annotations
+
+import os
+import sys
+from unittest.mock import MagicMock
+
+os.environ.setdefault("DATABASE_URL", "postgresql+psycopg://test:test@localhost:5432/test")
+
+_wp_mock = MagicMock()
+sys.modules.setdefault("weasyprint", _wp_mock)
+sys.modules.setdefault("weasyprint.CSS", _wp_mock)
+sys.modules.setdefault("weasyprint.HTML", _wp_mock)
+
+import pytest  # noqa: E402
+from fastapi import HTTPException  # noqa: E402
+
+from app.api.v1.admin import _assert_allowed_url  # noqa: E402
+from app.core.config import settings  # noqa: E402
+
+# ---------------------------------------------------------------------------
+# Absolute URL — host NOT in allowlist → 400
+# ---------------------------------------------------------------------------
+
+
+def test_absolute_ssrf_metadata_endpoint_blocked() -> None:
+    """http://169.254.169.254/... (AWS metadata) должен быть отклонён с 400."""
+    with pytest.raises(HTTPException) as exc_info:
+        _assert_allowed_url("http://169.254.169.254/latest/meta-data/")
+    assert exc_info.value.status_code == 400
+    assert "host not allowed" in exc_info.value.detail
+
+
+def test_absolute_localhost_blocked() -> None:
+    with pytest.raises(HTTPException) as exc_info:
+        _assert_allowed_url("http://localhost/internal/admin")
+    assert exc_info.value.status_code == 400
+
+
+def test_absolute_arbitrary_host_blocked() -> None:
+    with pytest.raises(HTTPException) as exc_info:
+        _assert_allowed_url("https://evil.example.com/scrape")
+    assert exc_info.value.status_code == 400
+
+
+# ---------------------------------------------------------------------------
+# Absolute URL — host IN allowlist → passes
+# ---------------------------------------------------------------------------
+
+
+def test_absolute_avito_allowed() -> None:
+    _assert_allowed_url("https://www.avito.ru/catalog/houses/some-slug/12345")
+
+
+def test_absolute_cian_allowed() -> None:
+    _assert_allowed_url("https://ekb.cian.ru/sale/flat/327830237/")
+
+
+def test_absolute_yandex_realty_allowed() -> None:
+    _assert_allowed_url("https://realty.yandex.ru/ekaterinburg/kupit/kvartira/123/")
+
+
+def test_absolute_n1_allowed() -> None:
+    _assert_allowed_url("https://ekaterinburg.n1.ru/kupit/kvartiry/vtorichka/")
+
+
+# ---------------------------------------------------------------------------
+# Relative path — no netloc → always passes
+# ---------------------------------------------------------------------------
+
+
+def test_relative_path_passes_the_guard() -> None:
+    """Путь без схемы/хоста (авито-стиль /ekaterinburg/...) разрешается."""
+    _assert_allowed_url("/ekaterinburg/kvartiry/prodam-123")
+
+
+def test_relative_path_house_catalog_passes() -> None:
+    _assert_allowed_url("/catalog/houses/ekb-slug/99999")
+
+
+def test_empty_string_passes() -> None:
+    """Пустая строка — нет netloc → guard не блокирует (upstream validation задача scraper'а)."""
+    _assert_allowed_url("")
+
+
+# ---------------------------------------------------------------------------
+# Allowlist integrity: verify expected hosts present
+# ---------------------------------------------------------------------------
+
+
+def test_allowlist_contains_required_scraper_hosts() -> None:
+    required = {
+        "www.avito.ru",
+        "ekb.cian.ru",
+        "www.cian.ru",
+        "realty.yandex.ru",
+        "ekaterinburg.n1.ru",
+    }
+    missing = required - settings.scrape_allowed_hosts
+    assert not missing, f"Missing hosts in scrape_allowed_hosts: {missing}"

tradein-mvp/docker-compose.prod.yml

@@ -58,8 +58,6 @@ services:
       # Yandex Geocoder API key — читается из backend/.env.runtime (env_file).
       # Раньше дублировалось здесь как ${YANDEX_GEOCODER_API_KEY:-} — убрано
       # т.к. environment: overrides env_file и при пустой host-env стирает значение.
-      # Защита /api/v1/admin/* — задаётся в .env.runtime. Пусто = открыто (dev).
-      ADMIN_TOKEN: "${ADMIN_TOKEN:-}"
       # GlitchTip DSN — мониторинг ошибок (#396). Пусто = выключено.
       GLITCHTIP_DSN: "${GLITCHTIP_DSN:-}"
       GENDESIGN_FDW_PASSWORD: "${GENDESIGN_FDW_PASSWORD:-}"