🛡️ Sentinel: [security improvement] 외부 API 호출 URL 스킴 검증 추가

.jules/sentinel.md

@@ -22,3 +22,7 @@
 **Vulnerability:** Server-Side Request Forgery (SSRF) / Local File Inclusion
 **Learning:** Functions that fetch URLs provided via user inputs (e.g., `wait_for_url` fetching `--backend-ready-url` in CI scripts) can inadvertently read local files if they do not validate the scheme. Python's `urllib.request.urlopen` supports `file://` schemes, allowing attackers to access arbitrary file contents from the host machine or sandbox if they can control the URL parameter.
 **Prevention:** Always validate URL inputs to restrict allowed schemes. Check that URLs explicitly start with `http://` or `https://` before fetching them with standard libraries like `urllib`.
+## 2026-07-04 - Fix unvalidated URL scheme in NOEMA_LLM_API_URL
+**Vulnerability:** urllib.request.urlopen in scripts/ci/noema_review_gate.py was calling the URL provided by NOEMA_LLM_API_URL environment variable without validating that it uses http:// or https:// scheme, potentially allowing file:// or other unexpected schemes to be used.
+**Learning:** In Python, urllib.request.urlopen will happily read local files if provided with a file:// scheme. This could lead to a Server-Side Request Forgery (SSRF) or local file read vulnerability if the environment variable is attacker-controlled or misconfigured. This expands our previous understanding of URL scheme validation to environment variables.
+**Prevention:** Always validate that URLs to be requested by urllib or other HTTP clients strictly start with http:// or https://, especially when the URL originates from environment variables or external configuration.

scripts/ci/noema_review_gate.py

@@ -267,6 +267,8 @@ def call_llm(repo: str, number: int, pr: dict[str, Any], diff: str, truncated: b
     if not api_url or not api_key:
         print("Noema LLM review unavailable: NOEMA_LLM_API_URL or NOEMA_LLM_API_KEY is not configured.")
         return None
+    if not (api_url.startswith("http://") or api_url.startswith("https://")):
+        raise ValueError(f"Invalid NOEMA_LLM_API_URL scheme: {api_url}")
 
     prompt = {
         "role": "user",
@@ -304,7 +306,7 @@ def call_llm(repo: str, number: int, pr: dict[str, Any], diff: str, truncated: b
         },
         method="POST",
     )
-    with urllib.request.urlopen(request, timeout=120) as response:
+    with urllib.request.urlopen(request, timeout=120) as response:  # nosec B310 - scheme validated above
         raw = response.read().decode("utf-8")
     data = json.loads(raw)
     content = (((data.get("choices") or [{}])[0].get("message") or {}).get("content") or "").strip()

tests/test_noema_review_gate.py

@@ -214,6 +214,11 @@ def fake_urlopen(request, timeout):
     assert seen["url"] == "https://llm.example.test/chat"
     assert seen["body"]["model"] == "review-model"
 
+    monkeypatch.setenv("NOEMA_LLM_API_URL", "file:///etc/passwd")
+    with pytest.raises(ValueError, match="Invalid NOEMA_LLM_API_URL scheme: file:///etc/passwd"):
+        noema.call_llm("owner/repo", 1, pr, "diff", True)
+
+    monkeypatch.setenv("NOEMA_LLM_API_URL", "https://llm.example.test/chat")
     monkeypatch.setattr(
         noema.urllib.request,
         "urlopen",