CVE-2025-4767

MEDIUM

defog-ai introspect <= 0.1.4 - Code Injection via Test Endpoint Input Model

Title source: llm
STIX 2.1

Description

A vulnerability was found in defog-ai introspect up to 0.1.4. It has been rated as critical. Affected by this issue is the function test_custom_tool of the file introspect/backend/integration_routes.py of the component Test Endpoint. The manipulation of the argument input_model leads to code injection. Attacking locally is a requirement. The exploit has been disclosed to the public and may be used.

References (5)

Core 5
Core References
Permissions Required, VDB Entry vdb-entry technical-description
https://vuldb.com/?id.309068
Permissions Required, VDB Entry signature permissions-required
https://vuldb.com/?ctiid.309068
Permissions Required, VDB Entry third-party-advisory
https://vuldb.com/?submit.571363

Scores

CVSS v3 5.3
EPSS 0.0018
EPSS Percentile 7.9%
Attack Vector LOCAL
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L

CISA SSVC

Vulnrichment
Exploitation poc
Automatable no
Technical Impact partial

Details

CWE
CWE-74 CWE-94
Status published
Products (5)
defog-ai/introspect 0.1.0
defog-ai/introspect 0.1.1
defog-ai/introspect 0.1.2
defog-ai/introspect 0.1.3
defog-ai/introspect 0.1.4
Published May 16, 2025
Tracked Since Feb 18, 2026