Red Hat Ansible Automation Platform EUVD-2026-85535

Requires an authenticated account of any privilege level on Red Hat Ansible Automation Platform 2 automation-controller (PR:L). The attacker sends a malformed value through the host_filter query parameter of the Host list REST API endpoint; HostList.list() catches a bare Exception and returns str(e) verbatim, so either a Django FieldError (via a bad relation lookup such as credential__search=x) or a PostgreSQL DataError (via a bad regex such as name__regex=[bad) reflects internal schema/error text. Limiting factors: valid credentials are mandatory (no unauthenticated path), impact is confined to information disclosure of the ORM relation graph and raw database error strings - no data modification, secret extraction, or code execution - and no user interaction is required.

This is a genuinely low-to-moderate priority information-disclosure flaw, and the CVSS 4.3 (MEDIUM) rating is appropriate. The RedHat-supplied vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N is internally consistent with the description: exploitation is network-reachable but requires an authenticated account (PR:L - 'any authenticated user'), needs no user interaction, and yields only limited confidentiality impact (C:L) with no integrity or availability effect. There is no EPSS score, no CISA KEV listing, and no public exploit code provided in the intelligence, so this is not evidence of widespread or automated exploitation. The impact ceiling is real but bounded: an attacker recovers the internal Django ORM relation graph (including reverse accessors) and raw PostgreSQL error strings - reconnaissance value that could seed a follow-on SQL/ORM-focused attack, but not direct access to credentials, secrets, or execution. No signals conflict: network vector aligns with the REST API surface (host_filter query parameter), and the PR:L requirement is consistent with the description's 'any authenticated user' phrasing. Treat as a defense-in-depth / recon-hardening fix rather than an emergency patch.

An authenticated low-privilege AAP user issues a crafted host_filter query against the automation-controller REST API - e.g. credential__search=x to force a Django FieldError that echoes the full Host model relation graph, or name__regex=[bad to force a PostgreSQL DataError whose raw text is reflected verbatim. The leaked schema and internal reverse accessors are then used to map the data model and plan targeted follow-on attacks. No public POC is referenced in the available intelligence, though the two documented filter primitives make reproduction straightforward.