Datasets:
Squidnutz
/
extractedtest

Dataset card Data Studio Files
xet
Community
1
text
stringlengths
0
152
number chosen by the attacker, even though you did not explicitly grant access to the
PRI to that caller. This problem can quickly cost you a whole lot of money.
There are several approaches to avoiding this problem. The first and easiest approach
is to always use strict pattern matching. If you know the length of extensions you are
expecting and expect only numeric extensions, use a strict numeric pattern match.
For example, this would work if you are expecting four-digit numeric extensions
only:
exten => _XXXX,1,Dial(PJSIP/otherserver/${EXTEN},30)
Another approach to mitigating dialplan injection vulnerabilities is by using the FIL
TER() dialplan function. Perhaps you would like to allow numeric extensions of any
length. FILTER() makes that easy to achieve safely:
exten => _X.,1,Set(SAFE_EXTEN=${FILTER(0-9A-F,${EXTEN})})
same => n,Dial(PJSIP/otherserver/${SAFE_EXTEN},30)
For more information about the syntax for the FILTER() dialplan function, see the
output of the core show function FILTER command at the Asterisk CLI.
A more comprehensive (but also complex) approach might be to have all dialed digits
validated by functions outside of your dialplan (for example, database queries that
validate the dialed string against user permissions, routing patterns, restriction tables,
and so forth). This is a powerful concept, but beyond the scope of this book.
Be wary of dialplan injection vulnerabilities. Use strict pattern
matching or use the FILTER() dialplan function to avoid these
problems.
372
|
Chapter 22: Security
Securing Asterisk Network APIs
To secure AGI, AMI, and ARI, you will need to carefully consider the following rec‐
ommended practices:
• Only allow connections directly to the API from localhost/127.0.0.1.
• Use an appropriate framework in between the Asterisk API and your client appli‐
cation, and handle connection security through the framework.
• Control access to the framework and the system through strict firewall rules.
Beyond that, the same sort of security rules and best practices apply that you would
follow in any mission-critical web application.
Other Risk Mitigation
There are other useful features in Asterisk that can be used to mitigate the risk of
attacks. The first is to use the permit and deny options to build access control lists
(ACLs) for privileged accounts. Consider a PBX that has SIP phones on a local net‐
work, but also accepts SIP calls from the public internet. Calls coming in over the
internet are only granted access to the main company menu, while local SIP phones
have the ability to make outbound calls that cost you money. In this case, it is a very
good idea to set ACLs to ensure that only devices on your local network can use the
accounts for the phones.
In your ps_endpoints table, the permit and deny options allow you to specify IP
addresses, but you can also point to a label in the /etc/asterisk/acl.conf file. In fact,
ACLs are accepted almost everywhere that connections to IP services are configured.
For example, another useful place for ACLs is in /etc/asterisk/manager.conf, to restrict
AMI accounts to the single host that is supposed to be using the manager interface.
ACLs can be defined in /etc/asterisk/acl.conf.
[named_acl_1]
deny=0.0.0.0/0.0.0.0
permit=10.1.1.50
permit=10.1.1.55
[named_acl_2] ; Named ACLs support IPv6, as well.
deny=::
permit=::1/128
[local_phones]
deny=0.0.0.0/0.0.0.0
permit=192.168.0.0/255.255.0.0
Once named ACLs have been defined in acl.conf, have Asterisk load them using the
reload acl command. Once loaded, they should be available via the Asterisk CLI:
Securing Asterisk Network APIs
|
373
*CLI> module reload acl
*CLI> acl show
acl
---
named_acl_1
named_acl_2
local_phones