Merge pull request 'fix(api): SSRF guard on ingest; whitelist PATCH /api/sources' (#46) from feat/ssrf-ingest-source-whitelist into master

Reviewed-on: #46

app/main.py

@@ -21,6 +21,7 @@ from datetime import datetime, timedelta, timezone
 from decimal import Decimal
 from pathlib import Path
 from typing import NoReturn
+from urllib.parse import urlparse
 from uuid import UUID
 
 import httpx
@@ -43,7 +44,7 @@ from schemas import (
     DashboardSummary, EntityCreate, EntityKind, EntityOut,
     EventCreate, EventOut, FireOut, NewsArticleOut, NewsMapItemOut,
     NewsSummaryOut, NewsTickerItemOut,
-    FeedSourceCreate, FeedSourceOut,
+    FeedSourceCreate, FeedSourceOut, FeedSourceUpdate,
     KeyOut, KeyValueIn,
     NewsModelsOut, SettingsIn, SettingsOut,
     SearchResult, SentimentSummary, SourceType,
@@ -51,7 +52,8 @@ from schemas import (
     GeofenceCreate, GeofenceUpdate,
 )
 from ingestor import ingest_event, fetch_and_process
-from sources import ingest_rss_feed, ingest_gdelt, ingest_earthquakes, ingest_social_signals
+from camera_scraper import is_public_url
+from sources import GDELT_API, ingest_rss_feed, ingest_gdelt, ingest_earthquakes, ingest_social_signals
 from fire_sources import ingest_fires
 from keystore import KeyFormatError, delete_key, list_keys, set_key
 from settings_store import SettingsError, get_app_settings, list_models, set_summary_model
@@ -366,18 +368,20 @@ async def create_source(payload: FeedSourceCreate):
 
 
 @app.patch("/api/sources/{source_id}")
-async def update_source(source_id: UUID, payload: dict):
+async def update_source(source_id: UUID, payload: FeedSourceUpdate):
-    """Update a feed source (e.g., toggle enabled)."""
+    """Update a feed source (name/url/config/enabled only)."""
+    values = payload.model_dump(exclude_unset=True)
     async with async_session() as session:
         row = (await session.execute(
             select(feed_sources).where(feed_sources.c.id == source_id)
         )).mappings().one_or_none()
         if not row:
             raise HTTPException(404, "Source not found")
+        if values:
             await session.execute(
                 feed_sources.update()
                 .where(feed_sources.c.id == source_id)
-            .values(**payload)
+                .values(**values)
             )
             await session.commit()
     return {"ok": True}
@@ -822,9 +826,15 @@ async def put_settings(payload: SettingsIn):
 
 # ── Ingestion Triggers ───────────────────────────────────────────────────
 
+def _require_public_url(url: str, field: str) -> None:
+    if not is_public_url(url):
+        raise HTTPException(400, f"{field} is not a public URL")
+
+
 @app.post("/api/ingest/rss")
 async def trigger_rss_ingest(feed_url: str, source_id: str | None = None):
     """Trigger RSS feed ingestion."""
+    _require_public_url(feed_url, "feed_url")
     count = await ingest_rss_feed(feed_url, source_id)
     return {"status": "ok", "items_ingested": count}
 
@@ -832,6 +842,10 @@ async def trigger_rss_ingest(feed_url: str, source_id: str | None = None):
 @app.post("/api/ingest/gdelt")
 async def trigger_gdelt_ingest(query: str = "", max_articles: int = 50):
     """Trigger GDELT data ingestion."""
+    _require_public_url(GDELT_API, "GDELT target")
+    parsed = urlparse(query)
+    if parsed.scheme in ("http", "https") and parsed.hostname:
+        _require_public_url(query, "query")
     count = await ingest_gdelt(query, max_articles)
     return {"status": "ok", "articles_ingested": count}
 

app/schemas.py

@@ -7,7 +7,7 @@ from enum import Enum
 from typing import Literal, Optional
 from uuid import UUID
 
-from pydantic import BaseModel, Field, field_validator
+from pydantic import BaseModel, ConfigDict, Field, field_validator
 
 
 # ─── Enums ───────────────────────────────────────────────────────────────
@@ -63,6 +63,17 @@ class FeedSourceCreate(BaseModel):
     config: Optional[dict] = None
 
 
+class FeedSourceUpdate(BaseModel):
+    """PATCH /api/sources/{id} — only these keys may be set."""
+
+    model_config = ConfigDict(extra="forbid")
+
+    name: Optional[str] = None
+    url: Optional[str] = None
+    config: Optional[dict] = None
+    enabled: Optional[bool] = None
+
+
 class FeedSourceOut(BaseModel):
     id: UUID
     name: str

tests/test_ingest_ssrf_and_sources.py

@@ -0,0 +1,67 @@
+"""SSRF guard on ingest triggers + PATCH /api/sources allowlist."""
+
+from __future__ import annotations
+
+import asyncio
+
+import httpx
+import pytest
+from pydantic import ValidationError
+
+from main import app
+
+BASE = "http://test"
+LINK_LOCAL_META = "http://169.254.169.254/latest/meta-data/"
+LOOPBACK = "http://127.0.0.1/secret"
+
+
+async def _req(method: str, path: str, **kw) -> httpx.Response:
+    transport = httpx.ASGITransport(app=app)
+    async with httpx.AsyncClient(transport=transport, base_url=BASE) as client:
+        return await client.request(method, path, **kw)
+
+
+def test_rss_ingest_rejects_link_local_metadata_url(monkeypatch):
+    called = {"n": 0}
+
+    async def _boom(*_a, **_k):
+        called["n"] += 1
+        raise AssertionError("ingest_rss_feed must not run for a private URL")
+
+    monkeypatch.setattr("main.ingest_rss_feed", _boom)
+    resp = asyncio.run(_req("POST", "/api/ingest/rss", params={"feed_url": LINK_LOCAL_META}))
+    assert resp.status_code == 400
+    assert called["n"] == 0
+
+
+def test_gdelt_ingest_rejects_private_query_url(monkeypatch):
+    called = {"n": 0}
+
+    async def _boom(*_a, **_k):
+        called["n"] += 1
+        raise AssertionError("ingest_gdelt must not run for a private URL query")
+
+    monkeypatch.setattr("main.ingest_gdelt", _boom)
+    resp = asyncio.run(_req("POST", "/api/ingest/gdelt", params={"query": LOOPBACK}))
+    assert resp.status_code == 400
+    assert called["n"] == 0
+
+
+def test_update_source_rejects_unknown_fields():
+    sid = "00000000-0000-0000-0000-000000000001"
+    resp = asyncio.run(_req("PATCH", f"/api/sources/{sid}", json={"enabled": True, "source_type": "rss"}))
+    assert resp.status_code == 422
+
+
+def test_feed_source_update_allowlist_only():
+    from schemas import FeedSourceUpdate
+
+    payload = FeedSourceUpdate(name="n", url="https://example.com/rss", config={"k": 1}, enabled=False)
+    assert payload.model_dump(exclude_unset=True) == {
+        "name": "n",
+        "url": "https://example.com/rss",
+        "config": {"k": 1},
+        "enabled": False,
+    }
+    with pytest.raises(ValidationError):
+        FeedSourceUpdate.model_validate({"enabled": True, "id": "00000000-0000-0000-0000-000000000001"})