Skip to content
DayZeroHub
Critical CVSS 10GHSA-gfv3-m769-5xm3CVE-2026-53998

Phoenix (Arize): Format String Injection in applyChatTemplate (CVE-2026-53998)

Arize-ai phoenix

Summary

Phoenix, Arize's AI observability and evaluation platform, has a format string injection vulnerability in the applyChatTemplate GraphQL query that can expose environment variables and internal state. The source rates it critical (CVSS 3.1: 10.0) and states it is fixed in version 15.7.0.

What happened

From version 13.0.0 up to but not including 15.7.0, the applyChatTemplate GraphQL query passed attacker-controlled F-string templates to FStringTemplateFormatter._format, and DotDict.__getattr__ did not prevent traversal through Python private attributes. A crafted field could traverse object internals and read process environment variables and internal module state. According to the source, the endpoint is unauthenticated on default installations, and any account can reach it when authentication is enabled. The associated identifiers are CVE-2026-53998 and GHSA-gfv3-m769-5xm3, with weaknesses CWE-134 and CWE-200.

Why it matters

According to the source, exposure of PHOENIX_SECRET can allow forgery of administrative session tokens and decryption of credentials protected by keys derived from that secret. The CVSS 3.1 score is 10.0 (vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:L, assigned by GitHub as CNA). The source does not indicate known exploitation, and the information available does not show this vulnerability in CISA's KEV catalog.

Affected products and versions

arize-phoenix (PyPI) / phoenix by Arize-ai: versions >= 13.0.0, < 15.7.0.

ProductVersionsFixed in
arize-phoenix>= 13.0.0, < 15.7.015.7.0
Arize-ai phoenix>= 13.0.0, < 15.7.0—

Action recommended by the source

The vendor states the issue is fixed in version 15.7.0. The source links the fix in pull request 13197 and in the arize-phoenix-v15.7.0 release of the Arize-ai/phoenix repository. The source does not state any other mitigations.

Fixed in: 15.7.0

Information taken from the source or the vendor. Verify it at the original link before acting.

Sources and references

Weaknesses (CWE): CWE-134, CWE-200

Relevant change history

  1. Published

    Advisory published.

Summary written with AI assistance from the source data. See the sources and references for complete information. View sources