Ansible Automation Platform CVE-2026-71465

Requires an authenticated AAP user (PR:L) with permission to run ad hoc commands, who must supply a limit/host-restriction value that begins with a '-' character so ansible-core misinterprets it as a CLI option. Exploitation is strongly limited today: the injected token displaces the required pattern positional, so only short-circuit flags (--version, --help) are reachable and no arbitrary option or command control is achievable. The description explicitly notes escalation 'would' only occur if a future ansible-core version defaults/makes optional the pattern positional - a non-default, hypothetical upstream change. AC:H reflects this reliance on specific parsing conditions.

This is a genuinely low-priority issue, and the signals are consistent rather than conflicting. The Red Hat-assigned CVSS 3.1 vector (AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:N, base 3.1) reflects an authenticated (PR:L), high-complexity (AC:H) argument-injection flaw with only Low integrity impact and no confidentiality or availability impact - an accurate characterization of a latent defect. No EPSS score, no CISA KEV listing, and no public POC were provided, so there is no evidence of exploitation interest. Critically, the description itself states the flaw is 'currently limited to short-circuit flags (--version, --help)' because the injected element displaces the required ansible pattern positional, meaning it cannot presently be turned into meaningful command control - it only becomes dangerous IF ansible-core were ever to make pattern default/optional. This is effectively a defense-in-depth / hardening fix for a code-quality asymmetry (RunAdHocCommand uses args.append(limit) while RunJob correctly uses args.extend(['-l', limit])). The 'Privilege Escalation' tag overstates the real-world impact; CWE-88 argument injection with I:L is more accurately an input-handling weakness. No key data is missing that would change the assessment - the low score, low CWE impact, and absence of exploitation signals all agree.

An authenticated Ansible Automation Platform user with permission to launch an ad hoc command supplies a limit value beginning with a dash (e.g. --version), which build_args() appends as a bare positional so ansible-core parses it as a CLI option rather than a host filter. In the current code path this only reaches short-circuit flags like --version/--help because the injected token displaces the mandatory pattern positional, so the practical effect is a malformed/aborted command rather than arbitrary option control. No public exploit code was identified at time of analysis.