From 4d60330338891e71b9668b14df35efd57e7b34e2 Mon Sep 17 00:00:00 2001 From: Tanmay9223 Date: Thu, 15 Oct 2026 00:53:46 +0530 Subject: [PATCH] fix(interface): Prevent argument injection in clone (closes #1134) Prevent Git argument injection and option smuggling when cloning repositories by inserting an end-of-options delimiter (--) before positional target arguments in clone_repository and rejecting leading-dash targets upfront in infer_target_type. Fix the following gaps: - Insert '--' separator before repo_url in git clone invocation - Reject targets and repo URLs starting with '-' in infer_target_type and clone_repository - Sanitize repo base names to strip leading dots/dashes and prevent directory traversal - Add comprehensive test coverage in tests/test_clone_repository.py --- strix/interface/utils.py | 25 +++++++++ tests/test_clone_repository.py | 94 ++++++++++++++++++++++++++++++++++ 2 files changed, 119 insertions(+) create mode 100644 tests/test_clone_repository.py diff --git a/strix/interface/utils.py b/strix/interface/utils.py index 7ab55d8e..15990b9e 100644 --- a/strix/interface/utils.py +++ b/strix/interface/utils.py @@ -1136,6 +1136,9 @@ def infer_target_type(target: str) -> tuple[str, dict[str, str]]: # noqa: PLR09 target = target.strip() + if target.startswith("-"): + raise ValueError(f"Invalid target: {target}. Targets cannot start with a dash.") + if target.startswith("git@"): return "repository", {"target_repo": target} @@ -1254,6 +1257,7 @@ def read_target_list_file(path_str: str) -> list[str]: def sanitize_name(name: str) -> str: sanitized = re.sub(r"[^A-Za-z0-9._-]", "-", name.strip()) + sanitized = sanitized.lstrip(".-") return sanitized or "target" @@ -1555,6 +1559,11 @@ def stage_api_specs(targets_info: list[dict[str, Any]], run_name: str) -> list[d def clone_repository(repo_url: str, run_name: str, dest_name: str | None = None) -> str: console = Console() + if not repo_url or repo_url.strip().startswith("-"): + raise ValueError( + f"Invalid repository URL: {repo_url}. Repository URL cannot start with a dash." + ) + git_executable = shutil.which("git") if git_executable is None: raise FileNotFoundError("Git executable not found in PATH") @@ -1567,8 +1576,23 @@ def clone_repository(repo_url: str, run_name: str, dest_name: str | None = None) else: repo_name = Path(repo_url).stem if repo_url.endswith(".git") else Path(repo_url).name + if ( + not repo_name + or repo_name in (".", "..") + or "/" in repo_name + or "\\" in repo_name + or repo_name.startswith("-") + ): + raise ValueError(f"Invalid repository destination name: {repo_name}") + clone_path = temp_dir / repo_name + if ( + not clone_path.resolve().is_relative_to(temp_dir.resolve()) + or clone_path.resolve() == temp_dir.resolve() + ): + raise ValueError(f"Invalid clone path outside temporary directory: {clone_path}") + if clone_path.exists(): shutil.rmtree(clone_path) @@ -1578,6 +1602,7 @@ def clone_repository(repo_url: str, run_name: str, dest_name: str | None = None) [ git_executable, "clone", + "--", repo_url, str(clone_path), ], diff --git a/tests/test_clone_repository.py b/tests/test_clone_repository.py new file mode 100644 index 00000000..0acdad4b --- /dev/null +++ b/tests/test_clone_repository.py @@ -0,0 +1,94 @@ +"""Tests for clone_repository, infer_target_type, and sanitization in strix.interface.utils.""" + +from __future__ import annotations + +from typing import TYPE_CHECKING +from unittest.mock import MagicMock, patch + +import pytest + +from strix.interface.utils import ( + clone_repository, + derive_repo_base_name, + infer_target_type, + sanitize_name, +) + + +if TYPE_CHECKING: + from pathlib import Path + + +def test_clone_repository_includes_double_dash_separator( + tmp_path: Path, monkeypatch: pytest.MonkeyPatch +) -> None: + monkeypatch.setattr("tempfile.gettempdir", lambda: str(tmp_path)) + monkeypatch.setattr("shutil.which", lambda cmd: "/usr/bin/git" if cmd == "git" else None) + + mock_run = MagicMock() + with patch("subprocess.run", mock_run): + res = clone_repository("https://github.com/example/test-repo.git", "run_123") + + expected_path = tmp_path / "strix_repos" / "run_123" / "test-repo" + assert res == str(expected_path.resolve()) + mock_run.assert_called_once_with( + [ + "/usr/bin/git", + "clone", + "--", + "https://github.com/example/test-repo.git", + str(expected_path), + ], + capture_output=True, + text=True, + check=True, + ) + + +def test_clone_repository_rejects_leading_dash_url() -> None: + with pytest.raises(ValueError, match="Repository URL cannot start with a dash"): + clone_repository("-u/upload-pack.git", "run_123") + + with pytest.raises(ValueError, match="Repository URL cannot start with a dash"): + clone_repository("--upload-pack=touch /tmp/pwn", "run_123") + + +def test_clone_repository_rejects_invalid_dest_names( + tmp_path: Path, monkeypatch: pytest.MonkeyPatch +) -> None: + monkeypatch.setattr("tempfile.gettempdir", lambda: str(tmp_path)) + monkeypatch.setattr("shutil.which", lambda cmd: "/usr/bin/git" if cmd == "git" else None) + + with pytest.raises(ValueError, match="Invalid repository destination name"): + clone_repository("https://example.com/repo.git", "run_123", dest_name="..") + + with pytest.raises(ValueError, match="Invalid repository destination name"): + clone_repository("https://example.com/repo.git", "run_123", dest_name=".") + + with pytest.raises(ValueError, match="Invalid repository destination name"): + clone_repository("https://example.com/repo.git", "run_123", dest_name="-dashed") + + with pytest.raises(ValueError, match="Invalid repository destination name"): + clone_repository("https://example.com/repo.git", "run_123", dest_name="sub/dir") + + +def test_infer_target_type_rejects_leading_dash() -> None: + with pytest.raises(ValueError, match="Targets cannot start with a dash"): + infer_target_type("-u/upload-pack.git") + + with pytest.raises(ValueError, match="Targets cannot start with a dash"): + infer_target_type("--target-flag") + + +def test_sanitize_name_strips_leading_dots_and_dashes() -> None: + assert sanitize_name("..") == "target" + assert sanitize_name(".") == "target" + assert sanitize_name("-dashed") == "dashed" + assert sanitize_name(".hidden") == "hidden" + assert sanitize_name("normal_repo-1") == "normal_repo-1" + + +def test_derive_repo_base_name_handles_dot_git() -> None: + assert derive_repo_base_name("..git") == "target" + assert derive_repo_base_name("https://github.com/foo/bar.git") == "bar" + assert derive_repo_base_name("git@github.com:foo/bar.git") == "bar"