Esri ArcGIS Server CVE-2025-57870

Exploitation requires the target ArcGIS Server (11.3, 11.4, or 11.5 on Windows, Linux, or Kubernetes) to be publishing Feature Services and to be network-reachable by the attacker; the injection is triggered through one specific ArcGIS Feature Service operation. Per the CVSS vector (PR:N/UI:N/AC:L) no authentication, no user interaction, and no non-default configuration are needed beyond having Feature Services enabled - a common, default use case for ArcGIS Server. The main limiting factor is exposure: an unpublished or network-isolated server without reachable Feature Services is not exploitable, and organizations already on a patched build are unaffected.

This is a genuine top-priority vulnerability, not a high-CVSS-but-low-real-risk case. The CVSS 3.1 vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H (10.0) is fully coherent with the description: a remote, unauthenticated attacker (PR:N/UI:N) sends network traffic (AV:N) with low complexity (AC:L) to a specific ArcGIS Feature Service operation and injects arbitrary SQL against the backing Enterprise Geodatabase. The scope change (S:C) is defensible because the injection crosses from the ArcGIS Server application into a distinct security authority - the geodatabase - yielding full confidentiality, integrity, and availability impact (read, modify, delete of enterprise geospatial data). Signals are incomplete: no EPSS probability, no CISA KEV listing, and no public POC were provided, so exploitation status is 'no public exploit identified at time of analysis.' The absence of EPSS/KEV does not lower the intrinsic severity - unauthenticated SQLi against a widely deployed enterprise server is exactly the profile that attracts rapid exploit development once a patch is diffed. No SSVC decision data was supplied. The one conflicting-metadata signal is the tag set (see data_quality_flags), which mislabels this as 'Authentication Bypass' and attributes it to 'Microsoft' rather than Esri; these are metadata errors, not evidence about the vulnerability itself.

An attacker locates an internet-facing or internally-reachable ArcGIS Server (versions 11.3-11.5) exposing Feature Services, then sends a crafted request to the specific vulnerable Feature Service operation with a malicious SQL payload embedded in an unsanitized parameter. Because no authentication is required, the injected SQL executes directly against the Enterprise Geodatabase, letting the attacker read, alter, or delete geospatial and business data. No public POC is referenced in the provided data, but the low attack complexity and unauthenticated network vector make weaponization straightforward once details circulate.