CVE-2026-105148

HighCVSS 7.3

Published: 04/10/2026(UTC)Updated: 05/10/2026(UTC)Translated: 04/10/2026(UTC)NVD NIST

Exploitation Probability (EPSS)

Low risk

0.28%

18th percentile - higher than 18% of all known CVEs

Summary

A server-side request forgery (SSRF) vulnerability was identified in SciPhi-AI R2R up to 3.6.6 in py/shared/abstractions/llm.py (Retrieval Completion API Endpoint component). Manipulation of the generation_config.api_base argument leads to SSRF. The exploit is publicly available and the attack can be launched remotely.

Risk Assessment

The vulnerability allows an attacker to force the server to make requests to internal services or resources, potentially leading to information disclosure or further attacks. The vendor's lack of response increases the risk.

Recommendation

Immediately validate and restrict the ability for users to set the api_base parameter, implement URL validation, and monitor outbound traffic from the server.

Other vulnerabilities in SciPhi-AI R2R

CVE-2026-105147High
A hard-coded credentials vulnerability was determined in SciPhi-AI R2R up to 3.6.6 in the JWT Secret Handler component. Manipulation of DEFAULT_BCRYPT_SECRET_KEY/DEFAULT_NACL_SECRET_KEY causes hard-coded credentials. The attack can be initiated remotely and the exploit is publicly disclosed.

Original NVD description (English source)

A vulnerability was identified in SciPhi-AI R2R up to 3.6.6. This vulnerability affects unknown code of the file py/shared/abstractions/llm.py of the component Retrieval Completion API Endpoint. Such manipulation of the argument generation_config.api_base leads to server-side request forgery. The attack can be launched remotely. The exploit is publicly available and might be used. The vendor was contacted early about this disclosure but did not respond in any way.

Vulnerability data from NVD (NIST) · CISA KEV · EPSS