diff --git a/.jules/sentinel.md b/.jules/sentinel.md index 9133bba1..a945caee 100644 --- a/.jules/sentinel.md +++ b/.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. diff --git a/scripts/ci/noema_review_gate.py b/scripts/ci/noema_review_gate.py index 1e4661b7..d57f6561 100644 --- a/scripts/ci/noema_review_gate.py +++ b/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() diff --git a/tests/test_noema_review_gate.py b/tests/test_noema_review_gate.py index 0b333ab3..aa9c5c04 100644 --- a/tests/test_noema_review_gate.py +++ b/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",