CM-57660-Remove PAT token from repository URL

Author: mateusz-sterczewski

cycode/cli/apps/report/sbom/repository_url/repository_url_command.py

@@ -8,6 +8,10 @@
 from cycode.cli.utils.get_api_client import get_report_cycode_client
 from cycode.cli.utils.progress_bar import SbomReportProgressBarSection
 from cycode.cli.utils.sentry import add_breadcrumb
+from cycode.cli.utils.url_utils import sanitize_repository_url
+from cycode.logger import get_logger
+
+logger = get_logger('Repository URL Command')
 
 
 def repository_url_command(
@@ -28,8 +32,13 @@ def repository_url_command(
     start_scan_time = time.time()
     report_execution_id = -1
 
+    # Sanitize repository URL to remove any embedded credentials/tokens before sending to API
+    sanitized_uri = sanitize_repository_url(uri)
+    if sanitized_uri != uri:
+        logger.debug('Sanitized repository URL to remove credentials')
+
     try:
-        report_execution = client.request_sbom_report_execution(report_parameters, repository_url=uri)
+        report_execution = client.request_sbom_report_execution(report_parameters, repository_url=sanitized_uri)
         report_execution_id = report_execution.id
 
         create_sbom_report(progress_bar, client, report_execution_id, output_file, output_format)

cycode/cli/apps/scan/remote_url_resolver.py

@@ -3,6 +3,7 @@
 from cycode.cli import consts
 from cycode.cli.utils.git_proxy import git_proxy
 from cycode.cli.utils.shell_executor import shell
+from cycode.cli.utils.url_utils import sanitize_repository_url
 from cycode.logger import get_logger
 
 logger = get_logger('Remote URL Resolver')
@@ -102,7 +103,11 @@ def _try_get_git_remote_url(path: str) -> Optional[str]:
         repo = git_proxy.get_repo(path, search_parent_directories=True)
         remote_url = repo.remotes[0].config_reader.get('url')
         logger.debug('Found Git remote URL, %s', {'remote_url': remote_url, 'repo_path': repo.working_dir})
-        return remote_url
+        # Sanitize URL to remove any embedded credentials/tokens before returning
+        sanitized_url = sanitize_repository_url(remote_url)
+        if sanitized_url != remote_url:
+            logger.debug('Sanitized repository URL to remove credentials')
+        return sanitized_url
     except Exception as e:
         logger.debug('Failed to get Git remote URL. Probably not a Git repository', exc_info=e)
         return None
@@ -124,7 +129,9 @@ def get_remote_url_scan_parameter(paths: tuple[str, ...]) -> Optional[str]:
         #  - len(paths)*2 Plastic SCM subprocess calls
         remote_url = _try_get_any_remote_url(path)
         if remote_url:
-            remote_urls.add(remote_url)
+            # URLs are already sanitized in _try_get_git_remote_url, but sanitize again as safety measure
+            sanitized_url = sanitize_repository_url(remote_url)
+            remote_urls.add(sanitized_url)
 
     if len(remote_urls) == 1:
         # we are resolving remote_url only if all paths belong to the same repo (identical remote URLs),

cycode/cli/utils/url_utils.py

@@ -0,0 +1,59 @@
+from typing import Optional
+from urllib.parse import urlparse, urlunparse
+
+from cycode.logger import get_logger
+
+logger = get_logger('URL Utils')
+
+
+def sanitize_repository_url(url: Optional[str]) -> Optional[str]:
+    """Remove credentials (username, password, tokens) from repository URL.
+
+    This function sanitizes repository URLs to prevent sending PAT tokens or other
+    credentials to the API. It handles both HTTP/HTTPS URLs with embedded credentials
+    and SSH URLs (which are returned as-is since they don't contain credentials in the URL).
+
+    Args:
+        url: Repository URL that may contain credentials (e.g., https://token@github.com/user/repo.git)
+
+    Returns:
+        Sanitized URL without credentials (e.g., https://github.com/user/repo.git), or None if input is None
+
+    Examples:
+        >>> sanitize_repository_url('https://token@github.com/user/repo.git')
+        'https://github.com/user/repo.git'
+        >>> sanitize_repository_url('https://user:token@github.com/user/repo.git')
+        'https://github.com/user/repo.git'
+        >>> sanitize_repository_url('git@github.com:user/repo.git')
+        'git@github.com:user/repo.git'
+        >>> sanitize_repository_url(None)
+        None
+    """
+    if not url:
+        return url
+
+    # Handle SSH URLs (git@host:path format) - no credentials to remove
+    if '@' in url and '://' not in url and url.startswith(('git@', 'ssh://')):
+        return url
+
+    try:
+        parsed = urlparse(url)
+        # Remove username and password from netloc
+        # Reconstruct URL without credentials
+        sanitized_netloc = parsed.hostname
+        if parsed.port:
+            sanitized_netloc = f'{sanitized_netloc}:{parsed.port}'
+
+        sanitized = urlunparse((
+            parsed.scheme,
+            sanitized_netloc,
+            parsed.path,
+            parsed.params,
+            parsed.query,
+            parsed.fragment,
+        ))
+        return sanitized
+    except Exception as e:
+        logger.debug('Failed to sanitize repository URL, returning original, %s', {'url': url, 'error': str(e)})
+        # If parsing fails, return original URL to avoid breaking functionality
+        return url

cycode/cyclient/report_client.py

@@ -6,8 +6,12 @@
 
 from cycode.cli.exceptions.custom_exceptions import CycodeError
 from cycode.cli.files_collector.models.in_memory_zip import InMemoryZip
+from cycode.cli.utils.url_utils import sanitize_repository_url
 from cycode.cyclient import models
 from cycode.cyclient.cycode_client_base import CycodeClientBase
+from cycode.logger import get_logger
+
+logger = get_logger('Report Client')
 
 
 @dataclasses.dataclass
@@ -49,7 +53,11 @@ def request_sbom_report_execution(
         # entity type required only for zipped-file
         request_data = {'report_parameters': params.to_json(without_entity_type=zip_file is None)}
         if repository_url:
-            request_data['repository_url'] = repository_url
+            # Sanitize repository URL to remove any embedded credentials/tokens before sending to API
+            sanitized_url = sanitize_repository_url(repository_url)
+            if sanitized_url != repository_url:
+                logger.debug('Sanitized repository URL to remove credentials')
+            request_data['repository_url'] = sanitized_url
 
         request_args = {
             'url_path': url_path,

tests/utils/test_url_utils.py

@@ -0,0 +1,82 @@
+import pytest
+
+from cycode.cli.utils.url_utils import sanitize_repository_url
+
+
+def test_sanitize_repository_url_with_token() -> None:
+    """Test that PAT tokens are removed from HTTPS URLs."""
+    url = 'https://token@github.com/user/repo.git'
+    expected = 'https://github.com/user/repo.git'
+    assert sanitize_repository_url(url) == expected
+
+
+def test_sanitize_repository_url_with_username_and_token() -> None:
+    """Test that username and token are removed from HTTPS URLs."""
+    url = 'https://user:token@github.com/user/repo.git'
+    expected = 'https://github.com/user/repo.git'
+    assert sanitize_repository_url(url) == expected
+
+
+def test_sanitize_repository_url_with_port() -> None:
+    """Test that URLs with ports are handled correctly."""
+    url = 'https://token@github.com:443/user/repo.git'
+    expected = 'https://github.com:443/user/repo.git'
+    assert sanitize_repository_url(url) == expected
+
+
+def test_sanitize_repository_url_ssh_format() -> None:
+    """Test that SSH URLs are returned as-is (no credentials in URL format)."""
+    url = 'git@github.com:user/repo.git'
+    assert sanitize_repository_url(url) == url
+
+
+def test_sanitize_repository_url_ssh_protocol() -> None:
+    """Test that ssh:// URLs are returned as-is."""
+    url = 'ssh://git@github.com/user/repo.git'
+    assert sanitize_repository_url(url) == url
+
+
+def test_sanitize_repository_url_no_credentials() -> None:
+    """Test that URLs without credentials are returned unchanged."""
+    url = 'https://github.com/user/repo.git'
+    assert sanitize_repository_url(url) == url
+
+
+def test_sanitize_repository_url_none() -> None:
+    """Test that None input returns None."""
+    assert sanitize_repository_url(None) is None
+
+
+def test_sanitize_repository_url_empty_string() -> None:
+    """Test that empty string is returned as-is."""
+    assert sanitize_repository_url('') == ''
+
+
+def test_sanitize_repository_url_gitlab() -> None:
+    """Test that GitLab URLs are sanitized correctly."""
+    url = 'https://oauth2:token@gitlab.com/user/repo.git'
+    expected = 'https://gitlab.com/user/repo.git'
+    assert sanitize_repository_url(url) == expected
+
+
+def test_sanitize_repository_url_bitbucket() -> None:
+    """Test that Bitbucket URLs are sanitized correctly."""
+    url = 'https://x-token-auth:token@bitbucket.org/user/repo.git'
+    expected = 'https://bitbucket.org/user/repo.git'
+    assert sanitize_repository_url(url) == expected
+
+
+def test_sanitize_repository_url_with_path_and_query() -> None:
+    """Test that URLs with paths, query params, and fragments are preserved."""
+    url = 'https://token@github.com/user/repo.git?ref=main#section'
+    expected = 'https://github.com/user/repo.git?ref=main#section'
+    assert sanitize_repository_url(url) == expected
+
+
+def test_sanitize_repository_url_invalid_url() -> None:
+    """Test that invalid URLs are returned as-is (graceful degradation)."""
+    # This should not raise an exception, but return the original
+    url = 'not-a-valid-url'
+    result = sanitize_repository_url(url)
+    # Should return original since parsing fails
+    assert result == url