GHSA-r54c-2xmf-2cf3: MS SWIFT Deserialization RCE Vulnerability
Noroxi assessment
Priority: Monitor (action score 0/100)
- The sources we track (CISA KEV, Exploit-DB, Metasploit, Nuclei, GitHub) show no public exploit code or exploitation in the wild.
- Anyone who can reach it over the internet can trigger it. No account or password is required. No special conditions are required; it is repeatable.
- No fixed version is recorded, but the references include a patch (commit or PR) link.
Low priority for now. See the fix and affected versions sections below.
Generated automatically from the record's data: CVSS (NVD and CNA), CISA KEV and SSVC, FIRST EPSS, exploit sources, fixed versions. How we calculate it
Primary sources: OSV
This appears to be a security vulnerability report describing a remote code execution (RCE) exploit in the ms-swift framework through malicious pickle deserialization in adapter model files. The vulnerability allows arbitrary command execution when loading specially crafted adapter models from ModelScope. This occurs when using machine torch version < 2.6.0, while ms-swift accepts torch version >= 2.0 **I. Detailed Description:** 1. Install ms-swift ``` pip install ms-swift -U ``` 2. Start web-ui ``` swift web-ui --lang en ``` 3. After startup, you can access [http://localhost:7860/](http://localhost:7860/) through your browser to see the launched fine-tuning framework program 4. Upload an adapter model repository (cyjhhh/lora_adapter_4_llama3) on ModelScope, where the lora/adapter_model.bin file is generated through the following code: ```python import torch, pickle, os class MaliciousPayload: def __reduce__(self): return (os.system, ("touch /tmp/malicious.txt",)) # Arbitrary command malicious_data = { "v_head.summary.weight": MaliciousPayload(), "v_head.summary.bias": torch.randn(10) } if __name__ == "__main__": with open("adapter_model.bin", "wb") as f: pickle.dump(malicious_data, f) ``` 5. First training submission: First, fill in the required parameters in the LLM Training interface, including Model id and Dataset Code, and configure the following in the Other params section of Advanced settings 6. Click Begin to submit. You can see the backend command running as follows 7. By reading the ms-swift source code, swift.llm.model.utils#safe_snapshot_download() and modelscope.hub.utils.utils#get_cache_dir(), we can see that adapters are downloaded locally to the path ~/.cache/modelscope. Therefore, the complete local path for the specified remote adapters after download is: ``` ~/.cache/modelscope/hub/models/cyjhhh/lora_adapter_4_llama3 ``` Wait for the first submission program until the adapters download is complete, then you can click "kill running task" on the page to terminate the first training 8. Second training submission, configure the page parameters as follows Click submit to see the backend command running as follows 9. After waiting for a while, you can see that torch.load() loaded the malicious adapter_model.bin file and successfully executed the command. Related execution information can also be seen in the log file corresponding to --logging_dir 10. Note (Prerequisites) Requires machine torch version < 2.6.0, while ms-swift accepts torch version >= 2.0 **II. Vulnerability Proof:** 1. Remote downloaded adapter malicious model: [[lora_adapter_4_llama3](https://www.modelscope.cn/models/cyjhhh/lora_adapter_4_llama3/files)](https://www.modelscope.cn/models/cyjhhh/lora_adapter_4_llama3/files) 2. For the second training submission, it's recommended to follow the parameters shown in the screenshots above for reproduction, as it will validate the target modules specified in the base model and adapter config. If they don't match, the program will terminate early. It's also recommended to select the same dataset content as shown in the screenshots 3. This report only reproduces RCE for one entry point (single path). In reality, there are more than one path in the code that can cause deserialization RCE **III. Fix Solution:** ``` SWIFT has disabled torch.load operations from 3.7 or later. ``` ## Author * Discovered by: [TencentAISec](https://github.com/TencentAISec) * Contact: *[security@tencent.com](mailto:security@tencent.com)*
- Published
- Jul 31, 2025
- Updated
- Aug 12, 2025
- EPSS
- —
- CWE
- CWE-502
Report tools
Action score
0
Low priority for now.
- CVSS
- 0 / 40 · —
- CISA KEV
- 0 / 30 · Not listed
- EPSS
- 0 / 30 · —
Get told when this record changes
- Not in KEV: no record of exploitation in the wild
- No public exploit code on record
- No fixed version yet
If any of these changes, we'll tell you in the daily email digest.
Is your version affected? Check itNoroxi analysis
What does it mean in your environment?
The assessment above is generated automatically from the record's data; it doesn't know your environment. Our team verifies on your systems whether this vulnerability is actually exploitable and which mitigation is enough.
We use this product, ask for helpAffected systems
Is your version affected?
Pick the product and type the version you run. It's compared with this record's version ranges in your browser; the version isn't sent anywhere.
| Vendor | Product | CPE |
|---|---|---|
| PyPI | ms-swift | pkg:pypi/ms-swift |
Package-level exposure
OSV and GitHub Advisory data: ecosystem, package and range. SBOM matching uses this table.
| Ecosystem | Package | Affected range | Fix |
|---|---|---|---|
| PyPI | ms-swift | up to and including 3.6.3 | — |
Remediation
Which version to upgrade to
Fix versions compiled from the vendor, package registries and Microsoft. Verify the vendor's note before upgrading.
No fix version is recorded for this entry. Check the references for vendor advisories.
Exploit status
No known public exploit
No public exploit has been observed yet. That doesn't mean you're safe, only that the bar is a little higher.
Research context
For pentesters and researchers: attack profile, score disagreement, timeline, patch commits, credits, variant and chain candidates, bug bounty scope. All derived from existing data; no exploit code.
Timeline
From publication to today: proof of concept, Metasploit module, CISA KEV and fix record. Dates are as reported by the sources.
No dated events beyond publication.
Patch and commit links
Commit, PR and diff links among the references. A starting point for patch-diffing and variant hunting; fixes, not exploits.
Not on record for this CVE: EPSS history · researcher credit · variant or chain candidate · public bounty program.
Technical details
Attack conditions
- Anyone who can reach it over the internet can trigger it.
- No account or password is required.
- No special conditions are required; it is repeatable.
If successful
- Confidentiality
- Integrity
- Availability
- Attack vector
- Network
- Attack complexity
- Low
- Privileges required
- None
- User interaction
- A
Weakness class (CWE)
CWE-502 · Deserialization of Untrusted DataCVSS vector
CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:P
osv
Root cause
Data received over the network is deserialized into the language’s native object format without verifying its source. Code paths inside the object can be triggered during deserialization.
The bug in code
The whole CWE-502 classA representative example of this vulnerability class. Not the vendor's source code; it shows the faulty pattern and its fix.
Vulnerable
payload = sock.recv(65536)
job = pickle.loads(payload)Fixed
payload = sock.recv(65536)
if not hmac.compare_digest(sign(payload), header_sig):
raise PermissionError("invalid signature")
job = JobSchema.validate(json.loads(payload))Attack context
MITRE CAPEC attack patterns and ATT&CK techniques for this weakness class (CWE). A starting point for detection rules and threat hunting.
Attack patterns (CAPEC)
ATT&CK techniques
—
Change log (1)
- CVSS5.5 → —
For records you follow, these changes also arrive as notifications. →
References
Show all references (2 more)Show less
Vendor advisories and official records. Exploit/PoC links are deliberately left out.