5.3

CVSS Score

3.1

-

CVSS Score

Basic Information

Concerned about an active attack path?

Talk to our security experts and see Miggo in action.

Miggo Vulnerability Database

→

CVE-2025-5148

CVE-2025-5148: FunAudioLLM InspireMusic deserialization vulnerability

A vulnerability was found in FunAudioLLM InspireMusic up to bf32364bcb0d136497ca69f9db622e9216b029dd. It has been classified as critical. Affected is the function load_state_dict of the file inspiremusic/cli/model.py of the component Pickle Data Handler. The manipulation leads to deserialization. An attack has to be approached locally. This product is using a rolling release to provide continious delivery. Therefore, no version details for affected nor updated releases are available. The name of the patch is 784cbf8dde2cf1456ff808aeba23177e1810e7a9. It is recommended to apply a patch to fix this issue.

(GitHub Advisory)

Miggo Vulnerability Database

→

CVE-2025-5148

CVE-2025-5148:

5.3

CVSS Score

3.1

-

CVSS Score

Basic Information

Is this CVE running in your environment?

Easily map the attack path and prioritize which CVEs are a threat to your organization

Validate Exposure

Technical Details

Package Name	Ecosystem	Vulnerable Versions	First Patched Version
inspiremusic	pip	<= 0.0.1.dev0

Technical Details

Vulnerability Intelligence
Miggo AI

Root Cause Analysis

The vulnerability exists in the load method of the InspireModel class, located in inspiremusic/cli/model.py. This method is responsible for loading model weights from files. The core issue is the use of torch.load for deserializing these model files without setting the weights_only=True parameter. By default, or when weights_only=False, torch.load uses Python's pickle module, which is known to be unsafe when loading data from untrusted sources as it can lead to arbitrary code execution. In this case, if an attacker can provide a maliciously crafted model file, the torch.load function will execute embedded code during deserialization. The deserialized (and potentially malicious) data is then passed to the load_state_dict method of the respective model components (self.llm and self.flow). The patch addresses this by adding weights_only=True to the torch.load calls, which restricts torch.load to only loading tensors and not arbitrary pickled objects, thus preventing the deserialization vulnerability.

Vulnerable functions

Only Mi**o us*rs **n s** t*is s**tion

Vulnerability Intelligence
Miggo AI

Unlock WAF rules for this CVE

Generate vendor-ready rules for the observed attack patterns, plus reasoning and safe deployment guidance

Get WAF rules

WAF Protection Rules

WAF Rule

W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.W** rul*s *v*il**l* *or Mi**o *ustom*rs only.

Reasoning

*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.*v*il**l* *or Mi**o *ustom*rs only.