Skip to content

OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https://www.npmjs.com/package/@openai/codex-security

fix(imports): validate Tenable repository URLs and CSV line numbers at capture (#2454)

Michael D'Angelo802058548586
diff --git a/examples/tenable-one/capture.py b/examples/tenable-one/capture.py
index 9eedf4cb6..8fd34475c 100644
--- a/examples/tenable-one/capture.py
+++ b/examples/tenable-one/capture.py
@@ -6,6 +6,7 @@ import json
import sys
from io import TextIOWrapper
from pathlib import Path
+from urllib.parse import urlsplit
from uuid import UUID
PLUGIN_SCRIPTS = Path(__file__).resolve().parents[2] / "plugins/codex-security/scripts"
@@ -69,8 +70,11 @@ def required_values(fields, row, context):
def validate_repository_url(repository_url):
_validate_remote(repository_url, "--repository-url")
- if not repository_url.lower().startswith("https://"):
+ parsed = urlsplit(repository_url)
+ if parsed.scheme != "https":
raise ValueError("--repository-url must be the repository's HTTPS URL")
+ if not parsed.path.strip("/"):
+ raise ValueError("--repository-url must include a repository path")
def read_asset(chunks, asset_id, *, exact=False):
@@ -109,6 +113,8 @@ def read_findings(chunks, asset_id, path_column, line_column):
if not line.isascii() or not line.isdecimal():
raise ValueError(f"{context}: {line_column} must be a nonnegative integer")
number = int(line)
+ if number > 2**53 - 1:
+ raise ValueError(f"{context}: {line_column} exceeds the maximum exact line number")
if number > 0:
if "code_file_path" not in finding:
raise ValueError(
diff --git a/examples/tenable-one/test_capture.py b/examples/tenable-one/test_capture.py
index 28e075a65..c390e0b20 100644
--- a/examples/tenable-one/test_capture.py
+++ b/examples/tenable-one/test_capture.py
@@ -151,6 +151,16 @@ class CaptureTests(unittest.TestCase):
chunk = self.csv("chunk", {**FINDING, "src": "lib/example.py", "pos": line})
self.capture(chunk, *LOCATIONS, error="must be a nonnegative integer")
+ def test_capture_only_writes_line_numbers_the_sdk_can_read_exactly(self):
+ row = {**FINDING, "src": "lib/example.py", "pos": "9007199254740992"}
+ self.capture(
+ self.csv("chunk", row), *LOCATIONS, error="exceeds the maximum exact line number"
+ )
+ self.assertFalse(self.output.exists())
+ row["pos"] = "9007199254740991"
+ result = self.capture(self.csv("chunk", row), *LOCATIONS)
+ self.assertEqual(result["findings"][0]["code_line_number"], 9007199254740991)
+
def test_existing_output_is_preserved(self):
self.output.write_bytes(b"previous")
self.capture(self.csv("chunk", FINDING), error="selection.json")
diff --git a/examples/tenable-one/test_download.py b/examples/tenable-one/test_download.py
index 85de130c4..04b1c80b7 100644
--- a/examples/tenable-one/test_download.py
+++ b/examples/tenable-one/test_download.py
@@ -46,6 +46,12 @@ class DownloadTests(unittest.TestCase):
self.request.side_effect = responses
download(ASSET, ids, URL, self.output, self.request, self.sleep)
+ def test_repository_host_alone_is_rejected_before_export(self):
+ for url in ("https://example.org", "https://example.org/"):
+ with self.subTest(url=url), self.assertRaisesRegex(ValueError, "repository path"):
+ download(ASSET, [FIRST], url, self.output, self.request)
+ self.request.assert_not_called()
+
def test_scoped_exports_poll_and_read_all_chunks(self):
# Give each job an ID distinct from its requested asset or finding ID.
self.run_download(