| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| JobJobEventsChildrenSummary view has no model/parent_model.
ModelAccessPermission.check_get_permissions() falls through
(returns True) for any authenticated user. The view uses
raw get_object_or_404(Job, pk) without DRF object-level
permission check. Zero-privilege user reads event tree
structure, event_processing_finished status, and enumerates
Job IDs platform-wide via 200/404 oracle. Sibling endpoint
/jobs/{id}/job_events/ correctly returns 403. |
| URLModificationMiddleware resolves named-URL lookups
against unfiltered Model.objects before RBAC. The 403→404
shim only rewrites 403 responses, leaving the pk=0 miss
path with a different 404 detail string. Differential
"Not found." vs "No <Model> matches..." reveals whether
a named resource (org, credential, inventory, host) exists
anywhere on the platform. Enables cross-tenant internal
hostname enumeration. |
| A container privilege escalation flaw was found in certain Ansible Automation Platform images. This issue arises from the /etc/passwd file being created with group-writable permissions during the build process. In certain conditions, an attacker who can execute commands within an affected container, even as a non-root user, can leverage their membership in the root group to modify the /etc/passwd file. This vulnerability allows an attacker to add a new user with any arbitrary UID, including UID 0, gaining full root privileges within the container. |
| A flaw was found in the AAP Controller's HashiCorp Vault credential plugin. The kubernetes_auth() function in awx_plugins/credentials/hashivault.py reads the controller pod's Kubernetes service account token and sends it to an attacker-controlled URL when a HashiCorp Vault Secret Lookup credential with kubernetes_role authentication is tested. An authenticated attacker with credential-creation privileges can exfiltrate the service account token, gaining Kubernetes API access to the control plane namespaces with full pod CRUD and secret read permissions, including database credentials and the Django SECRET_KEY. |
| A flaw was found in the Ansible Automation Platform automation-controller. When a
WorkflowJobTemplate is copied, the deep-copy permission sanitizer validates only the inventory,
unified_job_template, and credentials of each cloned node and fails to check the instance_groups
(and execution_environment and labels) that were preserved from the original. A user with
organization workflow-admin permission but no role on the referenced instance groups can copy a
workflow, become its administrator, and launch jobs pinned to instance groups they are not
authorized to use — including the control-plane instance group — bypassing the InstanceGroup
use_role boundary and causing attacker-influenced automation to run in the control-plane
execution context. |
| A flaw was found in Ansible Automation Platform's automation-controller. The custom
Credential Type environment-variable injector validates variable names against a
deny-list (an ANSIBLE_* prefix check plus a fixed ENV_BLOCKLIST) that omits
process-hijacking loader variables such as BASH_ENV, ENV, LD_PRELOAD, LD_LIBRARY_PATH,
PYTHONSTARTUP and GIT_SSH_COMMAND. Combined with the credential file injector, a
privileged user can write an attacker-controlled script into the execution environment
and point BASH_ENV at it, obtaining arbitrary code execution inside the
execution-environment container for any job that attaches a credential of that type. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. The setting that formats the log message emitted for API 4XX errors
is an administrator-controlled Python format-string template that is rendered
with a live user object as an argument. Because Python string formatting permits
attribute and item traversal on its arguments, an administrator can craft a
template that walks from the user object into the application settings and reads
the Django secret key and the database password. The formatted message is written
to a logger that can be forwarded to an external log aggregator, whose destination
is also administrator-controlled, allowing the secrets to be sent off the host. An
authenticated administrator can thereby obtain the master encryption key used to
protect all stored credentials and the database service password, enabling offline
decryption of every stored credential, forgery of user sessions, and direct
access to the controller database. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. The HTML view of job, ad hoc command, project update, and inventory
update standard output escapes HTML metacharacters but does not remove ANSI
terminal escape sequences before conversion to HTML. An ANSI OSC 8 hyperlink
sequence in the output is expanded into an HTML anchor whose href is not scheme-
filtered or escaped, so a low-privileged user who can produce output -- or an
external party whose data a playbook echoes -- can embed a javascript: link that
is rendered into a text/html response with no Content-Security-Policy. When a
higher-privileged user views the output page and clicks the link, attacker-
controlled JavaScript executes in their authenticated session, allowing actions
as that user up to full platform takeover. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. The Project scm_url field is not validated against values that
begin with a dash and is stored and passed verbatim to the git SCM module.
Because the module runs git ls-remote with the URL as a positional argument and
without a "--" separator, a git project URL such as "--upload-pack=<command>:x"
is interpreted by git as the --upload-pack option and executed via a shell. A
user with permission to create or modify a project in a single organization can
thereby execute arbitrary commands on the control-plane task pod, with output
reflected through the project update stdout endpoint, leading to cross-tenant
compromise and in-cluster lateral movement |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. Four debug views that trigger the internal task, dependency, and
workflow schedulers are configured to allow any user (including unauthenticated
clients) and are routed in production builds because their URL include is not
gated on the debug setting. An unauthenticated remote attacker can repeatedly
invoke these endpoints to acquire the cluster-wide scheduler advisory lock;
because the legitimate scheduler acquires the same lock without waiting, the
attacker causes real scheduler runs to be skipped, stalling job dispatch for
all tenants, while also consuming controller web workers. The debug root view
additionally discloses the list of debug endpoints to unauthenticated callers. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. The provisioning-callback secret (host_config_key) is exposed to
users holding only the read-level view_jobtemplate permission -- both in the
job template API representation and in the activity stream -- and the
provisioning callback endpoint trusts a client-supplied X-Forwarded-For
header to determine the calling host when the controller is deployed behind
the AAP gateway with an empty proxy allow-list. By reading the secret and
spoofing X-Forwarded-For to match any host in the job template's inventory, a
minimally privileged or unauthenticated remote attacker can launch the job
template against arbitrary managed hosts using the job template's credentials,
resulting in privilege escalation and remote code execution on managed hosts. |
| CopyAPIView (awx/awx/api/generics.py:873) sets permission_classes =
(IsAuthenticated,), so DRF's get_object() performs no object-level
RBAC. The get() handler (lines 988–991) explicitly guards with
request.user.can_access(obj._class_, 'read', obj) — but post()
(lines 1001–1010) does not. POST only checks:
can_access(model, 'add', create_kwargs_check)
can_access(model, 'copy_related', obj)
For JobTemplate, can_add (awx/awx/main/access.py:1465–1520) gates on
inventory.use_role + project.use_role +
execution_environment.read_role — resource-level roles that do not
imply read on the source JT — and can_copy_related (1522–1534) checks
only credentials.use_role. None of these imply the caller can read the
source JT. |
| A flaw was found in AWX. The container group pod_spec_override field uses an incomplete blocklist that only restricts automountServiceAccountToken, allowing injection of initContainers, serviceAccountName overrides, and projected service account token volumes. An AAP platform administrator can exploit this to escalate privileges to OpenShift namespace-level access and exfiltrate namespace secrets. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. The AWX_TASK_ENV setting accepts arbitrary environment variable
keys with no restriction, and its values are applied directly to the running
automation controller web and task processes rather than only to sandboxed
execution environments. A user with the system administrator role can set
variables such as REQUESTS_CA_BUNDLE, HTTPS_PROXY, SSLKEYLOGFILE, OPENSSL_CONF,
or LD_PRELOAD that reconfigure the control-plane process TLS trust store,
key-logging, OpenSSL engine, or dynamic linker. This enables silent interception
of the outbound TLS the control plane uses to fetch external secret-store
credentials and Red Hat subscription and Insights secrets, arbitrary file write
of TLS session keys, and potential native code execution in the control-plane
process, escalating an application administrator to compromise of the control
plane and all tenant secrets. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-controller. When a
custom credential type is created or updated, the injector configuration is validated by
rendering each attacker-supplied Jinja2 template synchronously inside the web request
worker, with no limit on iteration count, output size, or execution time. An
authenticated superuser can submit a credential type whose injector contains deeply
nested loops or large string operations, consuming CPU and wall-clock time in the web
worker until it is killed by the request timeout; issuing enough concurrent requests
places all web workers into a kill-and-respawn loop, making the Controller API and UI
unavailable for the duration of the attack. Additionally, certain template constructs
raise exception types that the validator does not handle, resulting in an unhandled
server error (HTTP 500) instead of a clean validation error. The rendering sandbox
prevents code execution, so there is no confidentiality or integrity impact. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. When attaching a source inventory to a constructed inventory through
the input_inventories relationship endpoint, the controller verifies only that
the requesting user can read the source inventory, rather than that they hold use
permission on it, unlike instance group attachment on the same access class. An
authenticated user who can administer a constructed inventory and has read-only
visibility of an inventory in another organization -- for example an
organization or system auditor -- can attach that foreign inventory as an input.
On synchronization the controller clones every host and host variable, including
secrets, into the attacker's inventory, and because the attacker administers the
constructed inventory they can run ad hoc commands against the cloned hosts,
resulting in cross-tenant disclosure of inventory data and secrets and code
execution against another tenant's managed hosts. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-controller.
The execute-permission check on a workflow job template node's unified job
template is skipped when the node's currently stored unified job template is
empty: the check inspects only the existing value, not the incoming one, and a
node can be created without a unified job template. An authenticated user who
holds admin permission on a single workflow job template can create an empty node
and then patch it to reference any job template, project, inventory source,
system job, or workflow on the platform -- including ones in other organizations
that they cannot otherwise read or launch. Running their own workflow then
executes the victim template with the victim's attached credentials, inventory
and project, resulting in cross-organization privilege escalation to arbitrary
automation execution. The patch response also discloses the victim template's
name and description. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-controller.
The host_filter query parameter on the hosts list API is parsed into a raw Django
ORM filter whose lookup path is validated only against a sensitive-field blocklist,
with no authorization check on the database relations it traverses. Because job
event and ad-hoc command output fields are not on that blocklist, an authenticated
user holding only the Read role on an inventory can construct filters that traverse
into the output of jobs they have no permission to view and use the returned host
count as a boolean oracle. Using regular-expression lookups, the attacker can
extract, character by character, the output (which routinely contains plaintext
credentials, tokens, and command results) of jobs and ad-hoc commands belonging to
other organizations, resulting in cross-tenant disclosure of job output |
| A flaw was found in Red Hat Ansible Automation Platform's automation-controller. A user
with the delegated Instance Group Admin role on a container group can set a free-form
pod specification override that is deep-merged, without filtering, into every job pod
launched on that group. The only hardening applied is forcing
automountServiceAccountToken to false, which does not remove an explicitly declared
projected serviceAccountToken volume or secret volume mounts. A non-superuser can
therefore cause the kubelet to mint a ServiceAccount token for the control-plane
automation-controller service account into the job container and mount arbitrary
secrets from the control-plane namespace, obtaining the control-plane service-account
identity on the Kubernetes API and cleartext control-plane secrets (including the
platform database password and Django SECRET_KEY), leading to full platform data
compromise. |
| A flaw was found in Red Hat Ansible Automation Platform's automation-
controller. The bulk job launch endpoint builds workflow job nodes from client
input using a serializer that leaves the node's job reference -- a field normally
set by the workflow task manager after it spawns a child job -- writable and does
not include it in the permission validation performed for the other node fields.
An authenticated user with permission to execute a single job template can submit
a bulk job launch whose node references the identifier of any unified job in any
organization, including jobs they cannot access. The node then exposes that job's
metadata, and cancelling the attacker's workflow cancels the referenced job
through the workflow cancellation cascade, without any per-job authorization
check. Repeated, this allows a low-privileged user to cancel running jobs,
project and inventory syncs, ad hoc commands, and system jobs across all
organizations, denying automation service platform-wide. |
