testssl.sh/doc/testssl.1.html
2025-07-20 18:24:17 +02:00

1273 lines
74 KiB
HTML
Raw Permalink Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8">
<title>testssl.sh</title>
<style type='text/css' media='all'>
body {
margin: 0;
padding: 0 5ex;
font-size: 14px;
font-family: monospace;
text-align: justify;
}
h2, h3, h4, h5, h6 {
color: #030201;
}
h2 {
font-size: 16px;
}
h3 {
font-size: 15px;
margin-left: 4ex;
}
p, pre, ul, ol, dl {
margin-left: 8ex;
}
code {
font-weight: bold;
color:#131211;
}
li p {
margin-left: 0;
}
pre > code {
display: block;
padding: 0;
white-space: pre-line;
}
ul > li > ul, ul > li > ol {
margin-left: 0;
}
</style>
</head>
<body>
<h2 id="name">NAME</h2>
<p>testssl.sh check encryption of SSL/TLS servers</p>
<h2 id="synopsis">SYNOPSIS</h2>
<p><code>testssl.sh [OPTIONS] &lt;URI&gt;</code>,
<code>testssl.sh [OPTIONS] --file &lt;FILE&gt;</code></p>
<p>or</p>
<p><code>testssl.sh [BANNER OPTIONS]</code></p>
<h2 id="description">DESCRIPTION</h2>
<p>testssl.sh is a free command line tool which checks a
servers service on any port for the support of TLS/SSL ciphers,
protocols as well as cryptographic flaws and much more.</p>
<p>The output rates findings by color (screen) or severity (file
output) so that you are able to tell whether something is good
or bad. The (screen) output has several sections in which
classes of checks are being performed. To ease readability on
the screen it aligns and indents the output properly.</p>
<p>Only you see the result. You also can use it internally on
your LAN. Except DNS lookups or unless you instruct testssl.sh
to check for revocation of certificates it doesnt use any other
hosts or even third parties for any test.</p>
<h2 id="requirements">REQUIREMENTS</h2>
<p>Testssl.sh is out of the box portable: it runs under any
Unix-like stack: Linux, *BSD, MacOS X, WSL=Windows Subsystem for
Linux, Cygwin and MSYS2. <code>bash</code> is a prerequisite,
also version 3 is still supported. Standard utilities like awk,
sed, tr and head are also needed. This can be of a BSD, System 5
or GNU flavor whereas grep from System V is not yet
supported.</p>
<p>Any OpenSSL or LibreSSL version is needed as a helper. Unlike
previous versions of testssl.sh almost every check is done via
(TCP) sockets. In addition statically linked OpenSSL binaries
for major operating systems are supplied in
<code>./bin/</code>.</p>
<h2 id="general">GENERAL</h2>
<p><code>testssl.sh URI</code> as the default invocation does
the so-called default run which does a number of checks and puts
out the results colorized (ANSI and termcap) on the screen. It
does every check listed below except <code>-E</code> which are
(order of appearance):</p>
<ol start="0" type="1">
<li><p>displays a banner (see below), does a DNS lookup also for
further IP addresses and does for the returned IP address a
reverse lookup. Last but not least a service check is being
done.</p></li>
<li><p>SSL/TLS protocol check</p></li>
<li><p>standard cipher categories</p></li>
<li><p>servers cipher preferences (server order?)</p></li>
<li><p>forward secrecy: ciphers and elliptical curves</p></li>
<li><p>server defaults (certificate info, TLS extensions,
session information)</p></li>
<li><p>HTTP header (if HTTP detected or being forced via
<code>--assume-http</code>)</p></li>
<li><p>vulnerabilities</p></li>
<li><p>testing each of 370 preconfigured ciphers</p></li>
<li><p>client simulation</p></li>
<li><p>rating</p></li>
</ol>
<p>If a target FQDN has multiple IPv4 and/or multiple IPv6
addresses, it scans all IPs with the specified options or using
the default run - unless specified otherwise, see
<code>--ip</code>, <code>-4</code> and <code>-6</code>. IPv6
connectivity is automagically checked. If theres noch such
thing you will see a banner <em>Testing all
<strong>IPv4</strong> addresses</em> and all IPv6 addresses will
appear in round brackets.</p>
<h2 id="options-and-parameters">OPTIONS AND PARAMETERS</h2>
<p>Options are either short or long options. Any long or short
option requiring a value can be called with or without an equal
sign. E.g.
<code>testssl.sh -t=smtp --wide --openssl=/usr/bin/openssl &lt;URI&gt;</code>
(short options with equal sign) is equivalent to
<code>testssl.sh --starttls smtp --wide --openssl /usr/bin/openssl &lt;URI&gt;</code>
(long option without equal sign). Some command line options can
also be preset via ENV variables.
<code>WIDE=true OPENSSL=/usr/bin/openssl testssl.sh --starttls=smtp &lt;URI&gt;</code>
would be the equivalent to the aforementioned examples.
Preference has the command line over any environment
variables.</p>
<p><code>&lt;URI&gt;</code> or <code>--file &lt;FILE&gt;</code>
always needs to be the last parameter.</p>
<h3 id="banner-options-standalone">BANNER OPTIONS
(standalone)</h3>
<p><code>--help</code> (or no arg) displays command line
help</p>
<p><code>-b, --banner</code> displays testssl.sh banner,
including license, usage conditions, version of testssl.sh,
detected openssl version, its path to it, # of ciphers of
openssl, its build date and the architecture.</p>
<p><code>-v, --version</code> same as before</p>
<p><code>-V [pattern], --local [pattern]</code> pretty print all
local ciphers supported by openssl version. If a pattern is
supplied it performs a match (ignore case) on any of the strings
supplied in the wide output, see below. The pattern will be
searched in the any of the columns: hexcode, cipher suite name
(OpenSSL or IANA), key exchange, encryption, bits. It does a
word pattern match for non-numbers, for number just a normal
match applies. Numbers here are defined as [0-9,A-F]. This means
(attention: catch) that the pattern CBC is matched as non-word,
but AES as word. This option also accepts
<code>--openssl=&lt;path_to_openssl&gt;</code>.</p>
<h3 id="input-parameters">INPUT PARAMETERS</h3>
<p><code>URI</code> can be a hostname, an IPv4 or IPv6 address
(restriction see below) or an URL. IPv6 addresses need to be in
square brackets. For any given parameter port 443 is assumed
unless specified by appending a colon and a port number. The
only preceding protocol specifier allowed is <code>https</code>.
You need to be aware that checks for an IP address might not hit
the vhost you want. DNS resolution (A/AAAA record) is being
performed unless you have an <code>/etc/hosts</code> entry for
the hostname.</p>
<p><code>--file &lt;fname&gt;</code> or the equivalent
<code>-iL &lt;fname&gt;</code> are mass testing options. Per
default it implicitly turns on <code>--warnings batch</code>,
unless warnings has been set to off before. In its first
incarnation the mass testing option reads command lines from
<code>fname</code>. <code>fname</code> consists of command lines
of testssl, one line per instance. Comments after <code>#</code>
are ignored, <code>EOF</code> signals the end of fname any
subsequent lines will be ignored too. You can also supply
additional options which will be inherited to each child,
e.g. When invoking
<code>testssl.sh --wide --log --file &lt;fname&gt;</code> . Each
single line in <code>fname</code> is parsed upon execution. If
theres a conflicting option and serial mass testing option is
being performed the check will be aborted at the time it occurs
and depending on the output option potentially leaving you with
an output file without footer. In parallel mode the mileage
varies, likely a line wont be scanned.</p>
<p>Alternatively <code>fname</code> can be in
<code>nmap</code>s grep(p)able output format
(<code>-oG</code>). Only open ports will be considered. Multiple
ports per line are allowed. The ports can be different and will
be tested by testssl.sh according to common practice in the
internet, i.e. if nmap shows in its output an open port 25,
automatically <code>-t smtp</code> will be added before the URI
whereas port 465 will be treated as a plain TLS/SSL port, not
requiring an STARTTLS SMTP handshake upfront. This is done by an
internal table which correlates nmaps open port detected to the
STARTTLS/plain text decision from testssl.sh.</p>
<p>Nmaps output always returns IP addresses and only if theres
a PTR DNS record available a hostname. As it is not checked by
nmap whether the hostname matches the IP (A or AAAA record),
testssl.sh does this automatically for you. If the A record of
the hostname matches the IP address, the hostname is used and
not the IP address. Please keep in mind that checks against an
IP address might not hit the vhost you maybe were aiming at and
thus it may lead to different results.</p>
<p>A typical internal conversion to testssl.sh file format from
nmaps grep(p)able format could look like:</p>
<pre><code> 10.10.12.16:443
10.10.12.16:1443
-t smtp host.example.com:25
host.example.com:443
host.example.com:631
-t ftp 10.10.12.11:21
10.10.12.11:8443</code></pre>
<p>Please note that <code>fname</code> has to be in Unix format.
DOS carriage returns wont be accepted. Instead of the command
line switch the environment variable FNAME will be honored
too.</p>
<p><code>--mode &lt;serial|parallel&gt;</code>. Mass testing to
be done serial (default) or parallel (<code>--parallel</code> is
shortcut for the latter, <code>--serial</code> is the opposite
option). Per default mass testing is being run in serial mode,
i.e. one line after the other is processed and invoked. The
variable <code>MASS_TESTING_MODE</code> can be defined to be
either equal <code>serial</code> or <code>parallel</code>.</p>
<p><code>--warnings &lt;batch|off&gt;</code>. The warnings
parameter determines how testssl.sh will deal with situations
where user input normally will be necessary. There are two
options. <code>batch</code> doesnt wait for a confirming
keypress when a client- or server-side problem is encountered.
As of 3.0 it just then terminates the particular scan. This is
automatically chosen for mass testing (<code>--file</code>).
<code>off</code> just skips the warning, the confirmation but
continues the scan, independent whether it makes sense or not.
Please note that there are conflicts where testssl.sh will still
ask for confirmation which are the ones which otherwise would
have a drastic impact on the results. Almost any other decision
will be made in the future as a best guess by testssl.sh. The
same can be achieved by setting the environment variable
<code>WARNINGS</code>.</p>
<p><code>--socket-timeout &lt;seconds&gt;</code> This is useful
for socket TCP connections to a node. If the node does not
complete a TCP handshake (e.g. because it is down or behind a
firewall or theres an IDS or a tarpit) testssl.sh may usually
hang for around 2 minutes or even much more. This parameter
instructs testssl.sh to wait at most <code>seconds</code> for
the handshake to complete before giving up. This option only
works if your OS has a timeout binary installed. SOCKET_TIMEOUT
is the corresponding environment variable. This doesnt work on
Macs out of the box.</p>
<p><code>--openssl-timeout &lt;seconds&gt;</code> This is
especially useful for all connects using openssl and practically
useful for mass testing. It avoids the openssl connect to hang
for ~2 minutes. The expected parameter <code>seconds</code>
instructs testssl.sh to wait before the openssl connect will be
terminated. The option is only available if your OS has a
timeout binary installed. As there are different implementations
of <code>timeout</code>: It automatically calls the binary with
the right parameters. OPENSSL_TIMEOUT is the equivalent
environment variable. This doesnt work on Macs out of the
box.</p>
<p><code>--basicauth &lt;user:pass&gt;</code> This can be set to
provide HTTP basic auth credentials which are used during checks
for security headers. BASICAUTH is the ENV variable you can use
instead.</p>
<p><code>--reqheader &lt;header&gt;</code> This can be used to
add additional HTTP request headers in the correct format
<code>Headername: headercontent</code>. This parameter can be
called multiple times if required. For example:
<code>--reqheader 'Proxy-Authorization: Basic dGVzdHNzbDpydWxlcw==' --reqheader 'ClientID: 0xDEADBEAF'</code>.
REQHEADER is the corresponding environment variable.</p>
<p><code>--mtls &lt;path_to_client_cert&gt;</code> This can be
set to provide a file containing a client certificatete and a
private key (not encrypted) in PEM format, which is used when a
mutual TLS authentication is required by the remote server. MTLS
is the equivalent environment variable.</p>
<h3 id="special-invocations">SPECIAL INVOCATIONS</h3>
<p><code>-t &lt;protocol&gt;, --starttls &lt;protocol&gt;</code>
does a default run against a STARTTLS enabled
<code>protocol</code>. <code>protocol</code> must be one of
<code>ftp</code>, <code>smtp</code>, <code>pop3</code>,
<code>imap</code>, <code>xmpp</code>, <code>sieve</code>,
<code>xmpp-server</code>, <code>telnet</code>,
<code>ldap</code>, <code>irc</code>, <code>lmtp</code>,
<code>nntp</code>, <code>postgres</code>, <code>mysql</code>.
For the latter four you need e.g. the supplied OpenSSL or
OpenSSL version 1.1.1. Please note: MongoDB doesnt offer a
STARTTLS connection, IRC currently only works with
<code>--ssl-native</code>. <code>irc</code> is WIP.</p>
<p><code>--xmpphost &lt;jabber_domain&gt;</code> is an
additional option for STARTTLS enabled XMPP: It expects the
jabber domain as a parameter. This is only needed if the domain
is different from the URI supplied.</p>
<p><code>--mx &lt;domain|host&gt;</code> tests all MX records
(STARTTLS on port 25) from high to low priority, one after the
other.</p>
<p><code>--ip &lt;ip&gt;</code> tests either the supplied IPv4
or IPv6 address instead of resolving host(s) in
<code>&lt;URI&gt;</code>. IPv6 addresses need to be supplied in
square brackets. <code>--ip=one</code> means: just test the
first A record DNS returns (useful for multiple IPs). If
<code>-6</code> and <code>--ip=one</code> was supplied an AAAA
record will be picked if available. The <code>--ip</code> option
might be also useful if you want to resolve the supplied
hostname to a different IP, similar as if you would edit
<code>/etc/hosts</code> or
<code>/c/Windows/System32/drivers/etc/hosts</code>.
<code>--ip=proxy</code> tries a DNS resolution via proxy.
<code>--ip=proxy</code> plus <code>--nodns=min</code> is useful
for situations with no local DNS as therell be no DNS timeouts
when trying to resolve CAA, TXT and MX records.</p>
<p><code>--proxy &lt;host&gt;:&lt;port&gt;</code> does ANY check
via the specified proxy. <code>--proxy=auto</code> inherits the
proxy setting from the environment. Any hostname supplied will
be resolved to the first A record, if it does not exist the AAAA
record is used. IPv4 and IPv6 addresses can be passed too, the
latter <em>also</em> with square bracket notation. Please note
that you need a newer OpenSSL or LibreSSL version for IPv6 proxy
functionality. In addition if you want lookups via proxy you can
specify <code>DNS_VIA_PROXY=true</code>. OCSP revocation
checking (<code>-S --phone-out</code>) is not supported by
OpenSSL via proxy. As supplying a proxy is an indicator for port
80 and 443 outgoing being blocked in your network an OCSP
revocation check wont be performed. However if
<code>IGN_OCSP_PROXY=true</code> has been supplied it will be
tried directly. Authentication to the proxy is not supported,
also no HTTPS or SOCKS proxy.</p>
<p><code>-6</code> scans only IPv6 addresses of the target.
Besides the OpenSSL binary supplied IPv6 is known to work with
vanilla OpenSSL &gt;= 1.1.0 and older versions &gt;=1.0.2 in
RHEL/CentOS/FC and Gentoo. Scans are somewhat in line with tools
like curl or wget, i.e. if theres an IPv6 address of the target
which can be reached, it just uses them. If you dont want this
behavior, you need to supply <code>-4.</code></p>
<p><code>-4</code> scans only IPv4 addresses of the target, IPv6
addresses of the target wont be scanned.</p>
<p><code>--ssl-native</code> Instead of using a mixture of bash
sockets and a few openssl s_client connects, testssl.sh uses the
latter (almost) only. This is faster but provides less accurate
results, especially for the client simulation and for cipher
support. For all checks you will see a warning if testssl.sh
cannot tell if a particular check cannot be performed. For some
checks however you might end up getting false negatives without
a warning. Thus it is not recommended to use. It should only be
used if you prefer speed over accuracy or you know that your
target has sufficient overlap with the protocols and cipher
provided by your openssl binary.</p>
<p><code>--openssl &lt;path_to_openssl&gt;</code> testssl.sh
tries first very hard to find the binary supplied (where the
tree of testssl.sh resides, from the directory where testssl.sh
has been started from, etc.). If all that doesnt work it falls
back to openssl supplied from the OS (<code>$PATH</code>). With
this option you can point testssl.sh to your binary of choice
and override any internal magic to find the openssl binary.
(Environment preset via
<code>OPENSSL=&lt;path_to_openssl&gt;</code>). Depending on your
test parameters it could be faster to pick the OpenSSL version
which has a bigger overlap in terms of ciphers protocols with
the target. Also, when testing a modern server, OpenSSL 3.X is
faster than older OpenSSL versions, or on MacOS 18, as opposed
to the provided LibreSSL version.</p>
<h3 id="tuning-options">TUNING OPTIONS</h3>
<p><code>--bugs</code> does some workarounds for buggy servers
like padding for old F5 devices. The option is passed as
<code>-bug</code> to openssl when needed, see
<code>s_client(1)</code>, environment preset via
<code>BUGS="-bugs"</code> (1x dash). For the socket part
testssl.sh has always workarounds in place to cope with broken
server implementations.</p>
<p><code>--assuming-http</code> testssl.sh normally does upfront
an application protocol detection. In cases where HTTP cannot be
automatically detected you may want to use this option. It
enforces testssl.sh not to skip HTTP specific tests (HTTP
header) and to run a browser based client simulation. Please
note that sometimes also the severity depends on the application
protocol, e.g. SHA1 signed certificates, the lack of any SAN
matches and some vulnerabilities will be punished harder when
checking a web server as opposed to a mail server.</p>
<p><code>-n, --nodns &lt;min|none&gt;</code> tells testssl.sh
which DNS lookups should be performed. <code>min</code> uses
only forward DNS resolution (A and AAAA record or MX record) and
skips CAA lookups and PTR records from the IP address back to a
DNS name. <code>none</code> performs no DNS lookups at all. For
the latter you either have to supply the IP address as a target,
to use <code>--ip</code> or have the IP address in
<code>/etc/hosts</code>. The use of the switch is only useful if
you either cant or are not willing to perform DNS lookups. The
latter can apply e.g. to some pentests. In general this option
could e.g. help you to avoid timeouts by DNS lookups.
<code>NODNS</code> is the environment variable for this.
<code>--nodns=min</code> plus <code>--ip=proxy</code> is useful
for situations with no local DNS as therell be no DNS timeouts
when trying to resolve CAA, TXT and MX records.</p>
<p><code>--sneaky</code> For HTTP header checks testssl.sh uses
normally the server friendly HTTP user agent
<code>TLS tester from ${URL}</code>. With this option your
traces are less verbose and a Firefox user agent is being used.
Be aware that it doesnt hide your activities. That is just not
possible (environment preset via <code>SNEAKY=true</code>).</p>
<p><code>--user-agent &lt;user agent&gt;</code> tells testssl.sh
to use the supplied HTTP user agent instead of the standard user
agent <code>TLS tester from ${URL}</code>.</p>
<p><code>--ids-friendly</code> is a switch which may help to get
a scan finished which otherwise would be blocked by a server
side IDS. This switch skips tests for the following
vulnerabilities: Heartbleed, CCS Injection, Ticketbleed and
ROBOT. The environment variable OFFENSIVE set to false will
achieve the same result. Please be advised that as an
alternative or as a general approach you can try to apply
evasion techniques by changing the variables USLEEP_SND and / or
USLEEP_REC and maybe MAX_WAITSOCK.</p>
<p><code>--phone-out</code> Checking for revoked certificates
via CRL and OCSP is not done per default. This switch instructs
testssl.sh to query external in a sense of the current run
URIs. By using this switch you acknowledge that the check might
have privacy issues, a download of several megabytes (CRL file)
may happen and there may be network connectivity problems while
contacting the endpoint which testssl.sh doesnt handle.
PHONE_OUT is the environment variable for this which needs to be
set to true if you want this.</p>
<p><code>--add-ca &lt;CAfile&gt;</code> enables you to add your
own CA(s) in PEM format for trust chain checks.
<code>CAfile</code> can be a directory containing files with a
.pem extension, a single file or multiple files as a comma
separated list of root CAs. Internally they will be added during
runtime to all CA stores. This is (only) useful for internal
hosts whose certificates are issued by internal CAs.
Alternatively ADDTL_CA_FILES is the environment variable for
this.</p>
<h3 id="single-check-options">SINGLE CHECK OPTIONS</h3>
<p>Any single check switch supplied as an argument prevents
testssl.sh from doing a default run. It just takes this and if
supplied other options and runs them - in the order they would
also appear in the default run.</p>
<p><code>-e, --each-cipher</code> checks each of the (currently
configured) 370 ciphers via openssl + sockets remotely on the
server and reports back the result in wide mode. If you want to
display each cipher tested you need to add
<code>--show-each</code>. Per default it lists the following
parameters: <code>hexcode</code>,
<code>OpenSSL cipher suite name</code>,
<code>key exchange</code>, <code>encryption bits</code>,
<code>IANA/RFC cipher suite name</code>. Please note the
<code>--mapping</code> parameter changes what cipher suite names
you will see here and at which position. Also please note that
the <strong>bit</strong> length for the encryption is shown and
not the <strong>security</strong> length, albeit itll be sorted
by the latter. For 3DES due to the Meet-in-the-Middle problem
the bit size of 168 bits is equivalent to the security size of
112 bits.</p>
<p><code>-E, --cipher-per-proto</code> is similar to
<code>-e, --each-cipher</code>. It checks each of the possible
ciphers, here: per protocol. If you want to display each cipher
tested you need to add <code>--show-each</code>. The output is
sorted by security strength, it lists the encryption bits
though.</p>
<p><code>-s, --std, --categories</code> tests certain lists of
cipher suites / cipher categories by strength.
(<code>--standard</code> is deprecated.) Those lists are
(<code>openssl ciphers $LIST</code>, $LIST from below:)</p>
<ul>
<li><code>NULL encryption ciphers</code>: NULL:eNULL</li>
<li><code>Anonymous NULL ciphers</code>: aNULL:ADH</li>
<li><code>Export ciphers</code> (w/o the preceding ones):
EXPORT:!ADH:!NULL</li>
<li><code>LOW</code> (64 Bit + DES ciphers, without EXPORT
ciphers):
LOW:DES:RC2:RC4:MD5:!ADH:!EXP:!NULL:!eNULL:!AECDH</li>
<li><code>3DES + IDEA ciphers</code>:
3DES:IDEA:!aNULL:!ADH:!MD5</li>
<li><code>Obsoleted CBC ciphers</code>:
HIGH:MEDIUM:AES:CAMELLIA:ARIA:!IDEA:!CHACHA20:!3DES:!RC2:!RC4:!AESCCM8:!AESCCM:!AESGCM:!ARIAGCM:!aNULL:!MD5</li>
<li><code>Strong ciphers with no FS</code> (AEAD):
AESGCM:CHACHA20:CamelliaGCM:AESCCM:ARIAGCM:!kEECDH:!kEDH:!kDHE:!kDHEPSK:!kECDHEPSK:!aNULL</li>
<li><code>Forward Secrecy strong ciphers</code> (AEAD):
AESGCM:CHACHA20:CamelliaGCM:AESCCM:ARIAGCM:!kPSK:!kRSAPSK:!kRSA:!kDH:!kECDH:!aNULL</li>
</ul>
<p><code>-f, --fs, --nsa, --forward-secrecy</code> Checks robust
forward secrecy key exchange. “Robust” means that ciphers having
intrinsic severe weaknesses like Null Authentication or
Encryption, 3DES and RC4 wont be considered here. There
shouldnt be the wrong impression that a secure key exchange has
been taking place and everything is fine when in reality the
encryption sucks. Also this section lists the available
elliptical curves and Diffie Hellman groups, as well as FFDHE
groups (TLS 1.2 and TLS 1.3).</p>
<p><code>-p, --protocols</code> checks TLS/SSL protocols SSLv2,
SSLv3, TLS 1.0 through TLS 1.3. And for HTTP also QUIC (HTTP/3),
SPDY (NPN) and ALPN (HTTP/2). For TLS 1.3 the final version and
several drafts (from 18 on) are tested. QUIC needs OpenSSL &gt;=
3.2 which can be automatically picked up when in
<code>/usr/bin/openssl</code> (or when defined environment
variable OPENSSL2). If a TLS-1.3-only host is encountered and
the openssl-bad version is used testssl.sh will e.g. for HTTP
header checks switch to <code>/usr/bin/openssl</code> (or when
defined via ENV to OPENSSL2). Also this will be tried for the
QUIC check.</p>
<p><code>-P, --server-preference, --preference</code> displays
the servers preferences: cipher order, with used openssl client:
negotiated protocol and cipher. If theres a cipher order
enforced by the server it displays it for each protocol
(openssl+sockets). If theres not, it displays instead which
ciphers from the server were picked with each protocol.</p>
<p><code>-S, --server_defaults</code> displays information from
the server hello(s):</p>
<ul>
<li>Available TLS extensions,</li>
<li>TLS ticket + session ID information/capabilities,</li>
<li>session resumption capabilities,</li>
<li>Time skew relative to localhost (most server implementations
return random values).</li>
<li>Several certificate information
<ul>
<li>signature algorithm,</li>
<li>key size,</li>
<li>key usage and extended key usage,</li>
<li>fingerprints and serial</li>
<li>Common Name (CN), Subject Alternative Name (SAN),
Issuer,</li>
<li>Trust via hostname + chain of trust against supplied
certificates</li>
<li>EV certificate detection</li>
<li>experimental “eTLS” detection</li>
<li>validity: start + end time, how many days to go (warning for
certificate lifetime &gt;=5 years)</li>
<li>revocation info (CRL, OCSP, OCSP stapling + must staple).
When <code>--phone-out</code> supplied it checks against the
certificate issuer whether the host certificate has been revoked
(plain OCSP, CRL).</li>
<li>displaying DNS Certification Authority Authorization
resource record</li>
<li>Certificate Transparency info (if provided by server).</li>
</ul></li>
</ul>
<p>For the trust chain check 5 certificate stores are provided.
If the test against one of the trust stores failed, the one is
being identified and the reason for the failure is displayed -
in addition the ones which succeeded are displayed too. You can
configure your own CA via ADDTL_CA_FILES, see section
<code>FILES</code> below. If the server provides no matching
record in Subject Alternative Name (SAN) but in Common Name
(CN), it will be indicated as this is deprecated. Also for
multiple server certificates are being checked for as well as
for the certificate reply to a non-SNI (Server Name Indication)
client hello to the IP address. Regarding the TLS clock skew: it
displays the time difference to the client. Only a few TLS
stacks nowadays still support this and return the local clock
<code>gmt_unix_time</code>, e.g. IIS, openssl &lt; 1.0.1f. In
addition to the HTTP date you could e.g. derive that there are
different hosts where your TLS and your HTTP request ended if
the time deltas differ significantly.</p>
<p><code>-x &lt;pattern&gt;, --single-cipher &lt;pattern&gt;</code>
tests matched <code>pattern</code> of ciphers against a server.
Patterns are similar to
<code>-V pattern , --local pattern</code>, see above about
matching.</p>
<p><code>-h, --header, --headers</code> if the service is HTTP
(either by detection or by enforcing via
<code>--assume-http</code>. It tests several HTTP headers
like</p>
<ul>
<li>HTTP Strict Transport Security (HSTS)</li>
<li>HTTP Public Key Pinning (HPKP)</li>
<li>Server banner</li>
<li>HTTP date+time</li>
<li>Server banner like Linux or other Unix vendor headers</li>
<li>Application banner (PHP, RoR, OWA, SharePoint, Wordpress,
etc)</li>
<li>Reverse proxy headers</li>
<li>Web server modules</li>
<li>IPv4 address in header</li>
<li>Cookie (including Secure/HTTPOnly flags)</li>
<li>Decodes BIG IP F5 non-encrypted cookies</li>
<li>Security headers (X-Frame-Options, X-XSS-Protection,
Expect-CT,… , CSP headers). Nonsense is not yet detected
here.</li>
</ul>
<p><code>-c, --client-simulation</code> This simulates a
handshake with a number of standard clients so that you can
figure out which client cannot or can connect to your site. For
the latter case the protocol, cipher and curve is displayed,
also if theres Forward Secrecy. testssl.sh uses a handselected
set of clients which are retrieved by the SSLlabs API. The
output is aligned in columns when combined with the
<code>--wide</code> option. If you want the full nine yards of
clients displayed use the environment variable ALL_CLIENTS.</p>
<p><code>-g, --grease</code> checks several server
implementation bugs like tolerance to size limitations and
GREASE, see RFC 8701. This check doesnt run per default.</p>
<h3 id="vulnerabilities">VULNERABILITIES</h3>
<p><code>-U, --vulnerable, --vulnerabilities</code> Just tests
all (of the following) vulnerabilities. The environment variable
<code>VULN_THRESHLD</code> determines after which value a
separate headline for each vulnerability is being displayed.
Default is <code>1</code> which means if you check for two
vulnerabilities, only the general headline for vulnerabilities
section is displayed in addition to the vulnerability and the
result. Otherwise each vulnerability or vulnerability section
gets its own headline in addition to the output of the name of
the vulnerability and test result. A vulnerability section is
comprised of more than one check, e.g. the renegotiation
vulnerability check has two checks, so has Logjam.</p>
<p><code>-H, --heartbleed</code> Checks for Heartbleed, a memory
leakage in openssl. Unless the server side doesnt support the
heartbeat extension it is likely that this check runs into a
timeout. The seconds to wait for a reply can be adjusted with
<code>HEARTBLEED_MAX_WAITSOCK</code>. 8 is the default.</p>
<p><code>-I, --ccs, --ccs-injection</code> Checks for CCS
Injection which is an openssl vulnerability. Sometimes also here
the check needs to wait for a reply. The predefined timeout of 5
seconds can be changed with the environment variable
<code>CCS_MAX_WAITSOCK</code>.</p>
<p><code>-T, --ticketbleed</code> Checks for Ticketbleed memory
leakage in BigIP loadbalancers.</p>
<p><code>--OP, --opossum</code> Checks for HTTP to HTTPS upgrade
vulnerability named Opossum.</p>
<p><code>--BB, --robot</code> Checks for vulnerability to ROBOT
/ (<em>Return Of Bleichenbachers Oracle Threat</em>)
attack.</p>
<p><code>--SI, --starttls-injection</code> Checks for STARTTLS
injection vulnerabilities (SMTP, IMAP, POP3 only).
<code>socat</code> and OpenSSL &gt;=1.1.0 is needed.</p>
<p><code>-R, --renegotiation</code> Tests renegotiation
vulnerabilities. Currently theres a check for <em>Secure
Renegotiation</em> and for <em>Secure Client-Initiated
Renegotiation</em>. Please be aware that vulnerable servers to
the latter can likely be DoSed very easily (HTTP). A check for
<em>Insecure Client-Initiated Renegotiation</em> is not yet
implemented.</p>
<p><code>-C, --compression, --crime</code> Checks for CRIME
(<em>Compression Ratio Info-leak Made Easy</em>) vulnerability
in TLS. CRIME in SPDY is not yet being checked for.</p>
<p><code>-B, --breach</code> Checks for BREACH (<em>Browser
Reconnaissance and Exfiltration via Adaptive Compression of
Hypertext</em>) vulnerability. As for this vulnerability HTTP
level compression is a prerequisite itll be not tested if HTTP
cannot be detected or the detection is not enforced via
<code>--assume-http</code>. Please note that only the URL
supplied (normally “/” ) is being tested.</p>
<p><code>-O, --poodle</code> Tests for SSL POODLE (<em>Padding
Oracle On Downgraded Legacy Encryption</em>) vulnerability. It
basically checks for the existence of CBC ciphers in SSLv3.</p>
<p><code>-Z, --tls-fallback</code> Checks TLS_FALLBACK_SCSV
mitigation. TLS_FALLBACK_SCSV is basically a ciphersuite
appended to the Client Hello trying to prevent protocol
downgrade attacks by a Man in the Middle.</p>
<p><code>-W, --sweet32</code> Checks for vulnerability to
SWEET32 by testing 64 bit block ciphers (3DES, RC2 and
IDEA).</p>
<p><code>-F, --freak</code> Checks for FREAK vulnerability
(<em>Factoring RSA Export Keys</em>) by testing for EXPORT RSA
ciphers</p>
<p><code>-D, --drown</code> Checks for DROWN vulnerability
(<em>Decrypting RSA with Obsolete and Weakened eNcryption</em>)
by checking whether the SSL 2 protocol is available at the
target. Please note that if you use the same RSA certificate
elsewhere you might be vulnerable too. testssl.sh doesnt check
for this but provides a helpful link @ censys.io which provides
this service.</p>
<p><code>-J, --logjam</code> Checks for LOGJAM vulnerability by
checking for DH EXPORT ciphers. It also checks for “common
primes” which are preconfigured DH keys. DH keys =&lt; 1024 Bit
will be penalized. Also FFDHE groups (TLS 1.2) will be displayed
here.</p>
<p><code>-A, --beast</code> Checks BEAST vulnerabilities in SSL
3 and TLS 1.0 by testing the usage of CBC ciphers.</p>
<p><code>-L, --lucky13</code> Checks for LUCKY13 vulnerability.
It checks for the presence of CBC ciphers in TLS versions 1.0 -
1.2.</p>
<p><code>-WS, --winshock</code> Checks for Winshock
vulnerability. It tests for the absence of a lot of ciphers,
some TLS extensions and ec curves which were introduced later in
Windows. In the end the server banner is being looked at.</p>
<p><code>--rc4, --appelbaum</code> Checks which RC4 stream
ciphers are being offered.</p>
<h3 id="output-options">OUTPUT OPTIONS</h3>
<p><code>-q, --quiet</code> Normally testssl.sh displays a
banner on stdout with several version information, usage rights
and a warning. This option suppresses it. Please note that by
choosing this option you acknowledge usage terms and the warning
normally appearing in the banner.</p>
<p><code>--wide</code> Except the “each cipher output” all tests
displays the single cipher name (scheme see below). This option
enables testssl.sh to display also for the following sections
the same output as for testing each ciphers: BEAST, FS, RC4. The
client simulation has also a wide mode. The difference here is
restricted to a column aligned output and a proper headline. The
environment variable <code>WIDE</code> can be used instead.</p>
<p><code>--mapping &lt;openssl|iana|no-openssl|no-iana&gt;</code></p>
<ul>
<li><code>openssl</code>: use the OpenSSL cipher suite name as
the primary name cipher suite name form (default),</li>
<li><code>iana</code>: use the IANA cipher suite name as the
primary name cipher suite name form.</li>
<li><code>no-openssl</code>: dont display the OpenSSL cipher
suite name, display IANA names only.</li>
<li><code>no-iana</code>: dont display the IANA cipher suite
name, display OpenSSL names only.</li>
</ul>
<p>Please note that in testssl.sh 3.0 you can still use
<code>rfc</code> instead of <code>iana</code> and
<code>no-rfc</code> instead of <code>no-iana</code> but itll
disappear after 3.0.</p>
<p><code>--show-each</code> This is an option for all wide modes
only: it displays all ciphers tested not only succeeded ones.
<code>SHOW_EACH_C</code> is your friend if you prefer to set
this via the shell environment.</p>
<p><code>--color &lt;0|1|2|3&gt;</code> determines the use of
colors on the screen and in the log file: <code>2</code> is the
default and makes use of ANSI and termcap escape codes on your
terminal. <code>1</code> just uses non-colored mark-up like
bold, italics, underline, reverse. <code>0</code> means no
mark-up at all = no escape codes. This is also what you want
when you want a log file without any escape codes.
<code>3</code> will color ciphers and EC according to an
internal (not yet perfect) rating. Setting the environment
variable <code>COLOR</code> to the value achieves the same
result. Please not that OpenBSD and early FreeBSD do not support
italics.</p>
<p><code>--colorblind</code> Swaps green and blue colors in the
output, so that this percentage of folks (up to 8% of males, see
https://en.wikipedia.org/wiki/Color_blindness) can distinguish
those findings better. <code>COLORBLIND</code> is the according
variable if you want to set this in the environment.</p>
<p><code>--debug &lt;0-6&gt;</code> This gives you additional
output on the screen (2-6), only useful for debugging.
<code>DEBUG</code> is the according environment variable which
you can use. There are six levels (0 is the default, thus it has
no effect):</p>
<ol type="1">
<li>screen output normal but leaves useful debug output in
<strong>/tmp/testssl.XXXXXX/</strong> . The info about the exact
directory is included in the screen output in the end of the
run.</li>
<li>lists more whats going on, status (high level) and
connection errors, a few general debug output</li>
<li>even slightly more info: hexdumps + other info</li>
<li>display bytes sent via sockets</li>
<li>display bytes received via sockets</li>
<li>whole 9 yards</li>
</ol>
<p><code>--disable-rating</code> disables rating. Rating
automatically gets disabled, to not give a wrong or misleading
grade, when not all required functions are executed (e.g when
checking for a single vulnerabilities).</p>
<h3 id="file-output-options">FILE OUTPUT OPTIONS</h3>
<p><code>--log, --logging</code> Logs stdout also to
<code>${NODE}-p${port}${YYYYMMDD-HHMM}.log</code> in current
working directory of the shell. Depending on the color output
option (see above) the output file will contain color and other
markup escape codes, unless you specify <code>--color 0</code>
too. <code>cat</code> and if properly configured
<code>less</code> will show the output properly formatted on
your terminal. The output shows a banner with the almost the
same information as on the screen. In addition it shows the
command line of the testssl.sh instance. Please note that the
resulting log file is formatted according to the width of your
screen while running testssl.sh. You can override the width with
the environment variable TERM_WIDTH.</p>
<p><code>--logfile &lt;logfile&gt;</code> or
<code>-oL &lt;logfile&gt;</code> Instead of the previous option
you may want to use this one if you want to log into a directory
or if you rather want to specify the log file name yourself. If
<code>logfile</code> is a directory the output will put into
<code>logfile/${NODE}-p${port}${YYYYMMDD-HHMM}.log</code>. If
<code>logfile</code> is a file it will use that file name, an
absolute path is also permitted here. LOGFILE is the variable
you need to set if you prefer to work environment variables
instead. Please note that the resulting log file is formatted
according to the width of your screen while running testssl.sh.
You can override the width with the environment variable
TERM_WIDTH.</p>
<p><code>--json</code> Logs additionally to JSON file
<code>${NODE}-p${port}${YYYYMMDD-HHMM}.json</code> in the
current working directory of the shell. The resulting JSON file
is opposed to <code>--json-pretty</code> flat which means each
section is self contained and has an identifier for each single
check, the hostname/IP address, the port, severity and the
finding. For vulnerabilities it may contain a CVE and CWE entry
too. The output doesnt contain a banner or a footer.</p>
<p><code>--jsonfile &lt;jsonfile&gt;</code> or
<code>-oj &lt;jsonfile&gt;</code> Instead of the previous option
you may want to use this one if you want to log the JSON out put
into a directory or if you rather want to specify the log file
name yourself. If <code>jsonfile</code> is a directory the
output will put into
<code>logfile/${NODE}-p${port}${YYYYMMDD-HHMM}.json</code>. If
<code>jsonfile</code> is a file it will use that file name, an
absolute path is also permitted here.</p>
<p><code>--json-pretty</code> Logs additionally to JSON file
<code>${NODE}-p${port}${YYYYMMDD-HHMM}.json</code> in the
current working directory of the shell. The resulting JSON file
is opposed to <code>--json</code> non-flat which means it is
structured. The structure contains a header similar to the
banner on the screen, including the command line, scan host,
openssl binary used, testssl version and epoch of the start
time. Then for every test section of testssl.sh it contains a
separate JSON object/section. Each finding has a key/value pair
identifier with the identifier for each single check, the
severity and the finding. For vulnerabilities it may contain a
CVE and CWE entry too. The footer lists the scan time in
seconds.</p>
<p><code>--jsonfile-pretty &lt;jsonfile&gt;</code> or
<code>-oJ &lt;jsonfile&gt;</code> Similar to the aforementioned
<code>--jsonfile</code> or <code>--logfile</code> it logs the
output in pretty JSON format (see <code>--json-pretty</code>)
into a file or a directory. For further explanation see
<code>--jsonfile</code> or <code>--logfile</code>.</p>
<p><code>--csv</code> Logs additionally to a CSV file
<code>${NODE}-p${port}${YYYYMMDD-HHMM}.csv</code> in the current
working directory of the shell. The output contains a header
with the keys, the values are the same as in the flat JSON
format (identifier for each single check, the hostname/IP
address, the port, severity, the finding and for vulnerabilities
a CVE and CWE number).</p>
<p><code>--csvfile &lt;csvfile&gt;</code> or
<code>-oC &lt;csvfile&gt;</code> Similar to the aforementioned
<code>--jsonfile</code> or <code>--logfile</code> it logs the
output in CSV format (see <code>--cvs</code>) additionally into
a file or a directory. For further explanation see
<code>--jsonfile</code> or <code>--logfile</code>.</p>
<p><code>--html</code> Logs additionally to an HTML file
<code>${NODE}-p${port}${YYYYMMDD-HHMM}.html</code> in the
current working directory of the shell. It contains a 1:1 output
of the console. In former versions there was a non-native option
to use “aha” (Ansi HTML Adapter: github.com/theZiz/aha) like
<code>testssl.sh [options] &lt;URI&gt; | aha &gt;output.html</code>.
This is not necessary anymore.</p>
<p><code>--htmlfile &lt;htmlfile&gt;</code> or
<code>-oH &lt;htmlfile&gt;</code> Similar to the aforementioned
<code>--jsonfile</code> or <code>--logfile</code> it logs the
output in HTML format (see <code>--html</code>) additionally
into a file or a directory. For further explanation see
<code>--jsonfile</code> or <code>--logfile</code>.</p>
<p><code>-oA &lt;filename&gt;</code> /
<code>--outFile &lt;filename&gt;</code> Similar to nmap it does
a file output to all available file formats: LOG, JSON pretty,
CSV, HTML. If the filename supplied is equal <code>auto</code>
the filename is automatically generated using <span
class="math inline"><em>N</em><em>O</em><em>D</em><em>E</em><em>p</em></span>{port}<span
class="math inline"><em>Y</em><em>Y</em><em>Y</em><em>Y</em><em>M</em><em>M</em><em>D</em><em>D</em><em>H</em><em>H</em><em>M</em><em>M</em>.</span>{EXT}
with the according extension. If a directory is provided all
output files will put into
<code>&lt;filename&gt;/${NODE}-p${port}${YYYYMMDD-HHMM}.{log,json,csv,html}</code>.</p>
<p><code>-oa &lt;filename&gt;</code> /
<code>--outfile &lt;filename&gt;</code> Does the same as the
previous option but uses flat JSON instead.</p>
<p><code>--hints</code> This option is not in use yet. This
option is meant to give hints how to fix a finding or at least a
help to improve something. GIVE_HINTS is the environment
variable for this.</p>
<p><code>--severity &lt;severity&gt;</code> For CSV and both
JSON outputs this will only add findings to the output file if a
severity is equal or higher than the <code>severity</code> value
specified. Allowed are
<code>&lt;LOW|MEDIUM|HIGH|CRITICAL&gt;</code>. WARN is another
level which translates to a client-side scanning error or
problem. Thus you will always see them in a file if they
occur.</p>
<p><code>--append</code> Normally, if an output file already
exists and it has a file size greater zero, testssl.sh will
prompt you to manually remove the file and exit with an error.
<code>--append</code> however will append to this file, without
a header. The environment variable APPEND does the same. Be
careful using this switch/variable. A complementary option which
overwrites an existing file doesnt exist per design.</p>
<p><code>--overwrite</code> Normally, if an output file already
exists and it has a file size greater zero, testssl.sh will not
allow you to overwrite this file. This option will do that
<strong>without any warning</strong>. The environment variable
OVERWRITE does the same. Be careful, you have been warned!</p>
<p><code>--outprefix &lt;fname_prefix&gt;</code> Prepend output
filename prefix <fname_prefix> before <code>${NODE}-</code>. You
can use as well the environment variable FNAME_PREFIX. Using
this any output files will be named
<code>&lt;fname_prefix&gt;-${NODE}-p${port}${YYYYMMDD-HHMM}.&lt;format&gt;</code>
when no file name of the respective output option was specified.
If you do not like the separator - you can as well supply a
<code>&lt;fname_prefix&gt;</code> ending in ., _ or ,. In
this case or if you already supplied - no additional - will
be appended to <code>&lt;fname_prefix&gt;</code>.</p>
<p>A few file output options can also be preset via environment
variables.</p>
<h3 id="color-ratings">COLOR RATINGS</h3>
<p>Testssl.sh makes use of (the eight) standard terminal colors.
The color scheme is as follows:</p>
<ul>
<li>light red: a critical finding</li>
<li>red: a high finding</li>
<li>brown: a medium finding</li>
<li>yellow: a low finding</li>
<li>green (blue if COLORBLIND is set): something which is either
in general a good thing or a negative result of a check which
otherwise results in a high finding</li>
<li>light green (light blue if COLORBLIND is set) : something
which is either in general a very good thing or a negative
result of a check which otherwise results in a critical
finding</li>
<li>no color at places where also a finding can be expected: a
finding on an info level</li>
<li>cyan: currently only used for <code>--show-each</code> or an
additional hint</li>
<li>magenta: signals a warning condition, e.g. either a local
lack of capabilities on the client side or another problem</li>
<li>light magenta: a fatal error which either requires strict
consent from the user to continue or a condition which leaves no
other choice for testssl.sh to quit</li>
</ul>
<p>What is labeled as “light” above appears as such on the
screen but is technically speaking “bold”. Besides
<code>--color=3</code> will color ciphers according to an
internal and rough rating.</p>
<p>Markup (without any color) is used in the following
manner:</p>
<ul>
<li>bold: for the name of the test</li>
<li>underline + bold: for the headline of each test section</li>
<li>underline: for a sub-headline</li>
<li>italics: for strings just reflecting a value read from the
server</li>
</ul>
<h3 id="tuning-via-env-variables-and-more-options">TUNING via
ENV variables and more options</h3>
<p>Except the environment variables mentioned above which can
replace command line options here a some which cannot be set
otherwise. Variables used for tuning are preset with reasonable
values. <em>There should be no reason to change them</em> unless
you use testssl.sh under special conditions.</p>
<ul>
<li>TERM_WIDTH is a variable which overrides the auto-determined
terminal width size. Setting this variable normally only makes
sense if you log the output to a file using the
<code>--log</code>, <code>--logfile</code> or <code>-oL</code>
option.</li>
<li>DEBUG_ALLINONE / SETX: when setting one of those to true
testssl.sh falls back to the standard bash behavior,
i.e. calling <code>bash -x testssl.sh</code> it displays the
bash debugging output not in an external file
<code>/tmp/testssl-&lt;XX&gt;.log</code></li>
<li>DEBUGTIME: Profiling option. When using bashs debug mode
and when this is set to true, it generates a separate text file
with epoch times in <code>/tmp/testssl-&lt;XX&gt;.time</code>.
They need to be concatenated by
<code>paste /tmp/testssl-&lt;XX&gt;.{time,log}</code> <!---
* FAST_SOCKET
* SHOW_SIGALGO
* FAST
--></li>
<li>EXPERIMENTAL=true is an option which is sometimes used in
the development process to make testing easier. In released
versions this has no effect.</li>
<li>ALL_CLIENTS=true runs a client simulation with <em>all</em>
(currently 126) clients when testing HTTP.</li>
<li>UNBRACKTD_IPV6: needs to be set to true for some old
versions of OpenSSL (like from Gentoo) which dont support
[bracketed] IPv6 addresses</li>
<li>NO_ENGINE: if you have problems with garbled output
containing the word engine you might want to set this to true.
It forces testssl.sh not try to configure openssls engine or a
non existing one from libressl</li>
<li>HEADER_MAXSLEEP: To wait how long before killing the process
to retrieve a service banner / HTTP header</li>
<li>MAX_WAITSOCK: It instructs testssl.sh to wait until the
specified time before declaring a socket connection dead. Dont
change this unless youre absolutely sure what youre doing.
Value is in seconds.</li>
<li>CCS_MAX_WAITSOCK Is the similar to above but applies only to
the CCS handshakes, for both of the two the two CCS payload.
Dont change this unless youre absolutely sure what youre
doing. Value is in seconds.</li>
<li>HEARTBLEED_MAX_WAITSOCK Is the similar to MAX_WAITSOCK but
applies only to the ServerHello after sending the Heartbleed
payload. Dont change this unless youre absolutely sure what
youre doing. Value is in seconds.</li>
<li>MEASURE_TIME_FILE For seldom cases when you dont want the
scan time to be included in the output you can set this to
false.</li>
<li>STARTTLS_SLEEP is per default set to 10 (seconds). Thats
the value testssl.sh waits for a string in the STARTTLS
handshake before giving up.</li>
<li>MAX_PARALLEL is the maximum number of tests to run in
parallel in parallel mass testing mode. The default value of 20
may be made larger on systems with faster processors.</li>
<li>MAX_WAIT_TEST is the maximum time (in seconds) to wait for a
single test in parallel mass testing mode to complete. The
default is 1200. <!---
* USLEEP_SND
* USLEEP_REC
--></li>
<li>HSTS_MIN is preset to 179 (days). If you want warnings
sooner or later for HTTP Strict Transport Security you can
change this.</li>
<li>HPKP_MIN is preset to 30 (days). If you want warnings sooner
or later for HTTP Public Key Pinning you can change this</li>
<li>DAYS2WARN1 is the first threshold when youll be warning of
a certificate expiration of a host, preset to 60 (days). For
Lets Encrypt this value will be divided internally by 2.</li>
<li>DAYS2WARN2 is the second threshold when youll be warning of
a certificate expiration of a host, preset to 30 (days). For
Lets Encrypt this value will be divided internally by 2.</li>
<li>TESTSSL_INSTALL_DIR is the derived installation directory of
testssl.sh. Relatively to that the <code>bin</code> and
mandatory <code>etc</code> directory will be looked for.</li>
<li>CA_BUNDLES_PATH: If you have an own set of CA bundles or you
want to point testssl.sh to a specific location of a CA bundle,
you can use this variable to set the directory which testssl.sh
will use. Please note that it overrides completely the builtin
path of testssl.sh which means that you will only test against
the bundles you point to. Also you might want to use
<code>~/utils/create_ca_hashes.sh</code> to create the hashes
for HPKP.</li>
<li>MAX_SOCKET_FAIL: A number which tells testssl.sh how often a
TCP socket connection may fail before the program gives up and
terminates. The default is 2. You can increase it to a higher
value if you frequently see a message like <em>Fatal error:
repeated openssl s_client connect problem, doesnt make sense to
continue</em>.</li>
<li>MAX_OSSL_FAIL: A number which tells testssl.sh how often an
OpenSSL s_client connect may fail before the program gives up
and terminates. The default is 2. You can increase it to a
higher value if you frequently see a message like <em>Fatal
error: repeated TCP connect problems, giving up</em>.</li>
<li>MAX_HEADER_FAIL: A number which tells testssl.sh how often a
HTTP GET request over OpenSSL may return an empty file before
the program gives up and terminates. The default is 3. Also here
you can increase the threshold when you spot messages like
<em>Fatal error: repeated HTTP header connect problems, doesnt
make sense to continue</em>.</li>
<li>OPENSSL2 can be used to supply an alternative openssl
version. This only makes sense if you want to amend the supplied
version in <code>bin/</code> which lacks TLS 1.3 support with a
version which doesn not and is not in
<code>/usr/bin/openssl</code>.</li>
<li>OSSL_SHORTCUT should be set to false when you run
interactively and dont want to switch automatically to
<code>/usr/bin/openssl</code> (<code>OPENSSL2</code>) if you
encounter a TLS 1.3-only host.</li>
</ul>
<h3 id="rating">RATING</h3>
<p>This program has a near-complete implementation of SSL Labss
<a
href="https://github.com/ssllabs/research/wiki/SSL-Server-Rating-Guide">SSL
Server Rating Guide</a>.</p>
<p>This is <em>not</em> a 100% reimplementation of the <a
href="https://www.ssllabs.com/ssltest/analyze.html">SSL Labs
SSL Server Test</a>, but an implementation of the above rating
specification, slight discrepancies may occur. Please note that
for now we stick to the SSL Labs rating as good as possible. We
are not responsible for their rating. Before filing issues
please inspect their Rating Guide.</p>
<p>Disclaimer: Having a good grade is <strong>NOT</strong>
necessarily equal to having good security! Dont start a
competition for the best grade, at least not without monitoring
the client handshakes and not without adding a portion of good
sense to it. Please note STARTTLS always results in a grade cap
to T. Anything else would lead to a false sense of security. Use
TLS, see also RFC 8314. The security of STARTTLS is always
client determined, i.e. checking the certificate which for SMTP
port 25 is often enough not the case. Also with DANE or MTA-STS
no one can test on the server side whether a client makes use if
it.</p>
<p>As of writing, these checks are missing:</p>
<ul>
<li>GOLDENDOODLE - should be graded <strong>F</strong> if
vulnerable</li>
<li>Insecure renegotiation - should be graded <strong>F</strong>
if vulnerable</li>
<li>Padding oracle in AES-NI CBC MAC check (CVE-2016-2107) -
should be graded <strong>F</strong> if vulnerable</li>
<li>Sleeping POODLE - should be graded <strong>F</strong> if
vulnerable</li>
<li>Zero Length Padding Oracle (CVE-2019-1559) - should be
graded <strong>F</strong> if vulnerable</li>
<li>Zombie POODLE - should be graded <strong>F</strong> if
vulnerable</li>
<li>All remaining old Symantec PKI certificates are distrusted -
should be graded <strong>T</strong></li>
<li>Symantec certificates issued before June 2016 are distrusted
- should be graded <strong>T</strong></li>
<li>Anonymous key exchange - should give <strong>0</strong>
points in <code>set_key_str_score()</code></li>
<li>Exportable key exchange - should give <strong>40</strong>
points in <code>set_key_str_score()</code></li>
<li>Weak key (Debian OpenSSL Flaw) - should give
<strong>0</strong> points in
<code>set_key_str_score()</code></li>
</ul>
<h4 id="implementing-new-grades-caps-or--warnings">Implementing
new grades caps or -warnings</h4>
<p>To implement a new grading cap, simply call the
<code>set_grade_cap()</code> function, with the grade and a
reason:</p>
<div class="sourceCode" id="cb2"><pre
class="sourceCode bash"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="ex">set_grade_cap</span> <span class="st">&quot;D&quot;</span> <span class="st">&quot;Vulnerable to documentation&quot;</span></span></code></pre></div>
<p>To implement a new grade warning, simply call the
<code>set_grade_warning()</code> function, with a message:</p>
<div class="sourceCode" id="cb3"><pre
class="sourceCode bash"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="ex">set_grade_warning</span> <span class="st">&quot;Documentation is always right&quot;</span></span></code></pre></div>
<h4
id="implementing-a-new-check-which-contains-grade-caps">Implementing
a new check which contains grade caps</h4>
<p>When implementing a new check (be it vulnerability or not)
that sets grade caps, the <code>set_rating_state()</code> has to
be updated (i.e. the <code>$do_mycheck</code> variable-name has
to be added to the loop, and <code>$nr_enabled</code>
if-statement has to be incremented)</p>
<p>The <code>set_rating_state()</code> automatically disables
rating, if all the required checks are <em>not</em> enabled.
This is to prevent giving out a misleading or wrong grade.</p>
<h4 id="implementing-a-new-revision">Implementing a new
revision</h4>
<p>When a new revision of the rating specification comes around,
the following has to be done:</p>
<ul>
<li>New grade caps has to be either:
<ol type="1">
<li>Added to the script wherever relevant, or</li>
<li>Added to the above list of missing checks (if above is not
possible)</li>
</ol></li>
<li>New grade warnings has to be added wherever relevant</li>
<li>The revision output in <code>run_rating()</code> function
has to updated</li>
</ul>
<h2 id="examples">EXAMPLES</h2>
<pre><code> testssl.sh testssl.sh</code></pre>
<p>does a default run on https://testssl.sh (protocols, standard
cipher lists, servers cipher preferences, forward secrecy,
server defaults, vulnerabilities, client simulation, and
rating.</p>
<pre><code> testssl.sh testssl.net:443</code></pre>
<p>does the same default run as above with the subtle difference
that testssl.net has two IPv4 addresses. Both are tested.</p>
<pre><code> testssl.sh --ip=one --wide https://testssl.net:443</code></pre>
<p>does the same checks as above, with the difference that one
IP address is being picked randomly. Displayed is everything
where possible in wide format.</p>
<pre><code> testssl.sh -6 https://testssl.net</code></pre>
<p>As opposed to the first example it also tests the IPv6 part
supposed you have an IPv6 network and your openssl supports IPv6
(see above).</p>
<pre><code> testssl.sh -t smtp smtp.gmail.com:25</code></pre>
<p>Checks are done via a STARTTLS handshake on the plain text
port 25. It checks every IP on smtp.gmail.com.</p>
<pre><code> testssl.sh --starttls=imap imap.gmx.net:143</code></pre>
<p>does the same on the plain text IMAP port.</p>
<p>Please note that for plain TLS-encrypted ports you must not
specify the protocol option when no STARTTLS handshake is
offered: <code>testssl.sh smtp.gmail.com:465</code> just checks
the encryption on the SMTPS port,
<code>testssl.sh imap.gmx.net:993</code> on the IMAPS port. Also
MongoDB which provides TLS support without STARTTLS can be
tested directly.</p>
<h2 id="rfcs-and-other-standards">RFCs and other standards</h2>
<ul>
<li>RFC 2246: The TLS Protocol Version 1.0</li>
<li>RFC 2595: Using TLS with IMAP, POP3 and ACAP</li>
<li>RFC 2817: Upgrading to TLS Within HTTP/1.1</li>
<li>RFC 2818: HTTP Over TLS</li>
<li>RFC 2830: Lightweight Directory Access Protocol (v3):
Extension for Transport Layer Security</li>
<li>RFC 3207: SMTP Service Extension for Secure SMTP over
Transport Layer Security</li>
<li>RFC 3501: INTERNET MESSAGE ACCESS PROTOCOL - VERSION
4rev1</li>
<li>RFC 4346: The Transport Layer Security (TLS) Protocol
Version 1.1</li>
<li>RFC 4366: Transport Layer Security (TLS) Extensions</li>
<li>RFC 4492: Elliptic Curve Cryptography (ECC) Cipher Suites
for Transport Layer Security (TLS)</li>
<li>RFC 5077: Transport Layer Security (TLS) Session
Resumption</li>
<li>RFC 5246: The Transport Layer Security (TLS) Protocol
Version 1.2</li>
<li>RFC 5280: Internet X.509 Public Key Infrastructure
Certificate and Certificate Revocation List (CRL) Profile</li>
<li>RFC 5321: Simple Mail Transfer Protocol</li>
<li>RFC 5746: Transport Layer Security (TLS) Renegotiation
Indication Extension</li>
<li>RFC 5804: A Protocol for Remotely Managing Sieve
Scripts</li>
<li>RFC 6066: Transport Layer Security (TLS) Extensions:
Extension Definitions</li>
<li>RFC 6101: The Secure Sockets Layer (SSL) Protocol Version
3.0</li>
<li>RFC 6120: Extensible Messaging and Presence Protocol (XMPP):
Core</li>
<li>RFC 6125: Domain-Based Application Service Identity
[..]</li>
<li>RFC 6797: HTTP Strict Transport Security (HSTS)</li>
<li>RFC 6961: The Transport Layer Security (TLS) Multiple
Certificate Status Request Extension</li>
<li>RFC 7469: Public Key Pinning Extension for HTTP (HPKP)</li>
<li>RFC 7507: TLS Fallback Signaling Cipher Suite Value (SCSV)
for Preventing Protocol Downgrade Attacks</li>
<li>RFC 7627: Transport Layer Security (TLS) Session Hash and
Extended Master Secret Extension</li>
<li>RFC 7633: X.509v3 Transport Layer Security (TLS) Feature
Extension</li>
<li>RFC 7465: Prohibiting RC4 Cipher Suites</li>
<li>RFC 7685: A Transport Layer Security (TLS) ClientHello
Padding Extension</li>
<li>RFC 7905: ChaCha20-Poly1305 Cipher Suites for Transport
Layer Security (TLS)</li>
<li>RFC 7919: Negotiated Finite Field Diffie-Hellman Ephemeral
Parameters for Transport Layer Security</li>
<li>RFC 8143: Using Transport Layer Security (TLS) with Network
News Transfer Protocol (NNTP)</li>
<li>RFC 8446: The Transport Layer Security (TLS) Protocol
Version 1.3</li>
<li>RFC 8701: Applying Generate Random Extensions And Sustain
Extensibility (GREASE) to TLS Extensibility</li>
<li>RFC 9000: QUIC: A UDP-Based Multiplexed and Secure
Transport</li>
<li>W3C CSP: Content Security Policy Level 1-3</li>
<li>TLSWG Draft: The Transport Layer Security (TLS) Protocol
Version 1.3</li>
</ul>
<h2 id="exit-status">EXIT STATUS</h2>
<ul>
<li>0 testssl.sh finished successfully without errors and
without ambiguous results</li>
<li>1 testssl.sh has encountered exactly one ambiguous situation
or an error during run</li>
<li>1+n same as previous. The errors or ambiguous results are
added, also per IP.</li>
<li>50-200 reserved for returning a vulnerability scoring for
system monitoring or a CI tools</li>
<li>242 (ERR_CHILD) Child received a signal from master</li>
<li>244 (ERR_RESOURCE) Resources testssl.sh needs couldnt be
read</li>
<li>245 (ERR_CLUELESS) Weird state, either though user options
or testssl.sh</li>
<li>246 (ERR_CONNECT) Connectivity problem</li>
<li>247 (ERR_DNSLOOKUP) Problem with resolving IP addresses or
names</li>
<li>248 (ERR_OTHERCLIENT) Other client problem</li>
<li>249 (ERR_DNSBIN) Problem with DNS lookup binaries</li>
<li>250 (ERR_OSSLBIN) Problem with OpenSSL binary</li>
<li>251 (ERR_NOSUPPORT) Feature requested is not supported</li>
<li>252 (ERR_FNAMEPARSE) Input file couldnt be parsed</li>
<li>253 (ERR_FCREATE) Output file couldnt be created</li>
<li>254 (ERR_CMDLINE) Cmd line couldnt be parsed</li>
<li>255 (ERR_BASH) Bash version incorrect</li>
</ul>
<h2 id="files">FILES</h2>
<p><strong>etc/*pem</strong> are the certificate stores from
Apple, Linux, Mozilla Firefox, Windows and Java.</p>
<p><strong>etc/client-simulation.txt</strong> contains client
simulation data.</p>
<p><strong>etc/cipher-mapping.txt</strong> provides a mandatory
file with mapping from OpenSSL cipher suites names to the ones
from IANA / used in the RFCs.</p>
<p><strong>etc/tls_data.txt</strong> provides a mandatory file
for ciphers (bash sockets) and key material.</p>
<h2 id="authors">AUTHORS</h2>
<p>Developed by Dirk Wetter, David Cooper and many others, see
CREDITS.md .</p>
<h2 id="copyright">COPYRIGHT</h2>
<p>Copyright © 2012 Dirk Wetter. License GPLv2: Free Software
Foundation, Inc. This is free software: you are free to change
and redistribute it under the terms of the license, see
LICENSE.</p>
<p>Attribution is important for the future of this project -
also in the internet. Thus if youre offering a scanner based on
testssl.sh as a public and/or paid service in the internet you
are strongly encouraged to mention to your audience that youre
using this program and where to get this program from. That
helps us to get bugfixes, other feedback and more
contributions.</p>
<p>Usage WITHOUT ANY WARRANTY. USE at your OWN RISK!</p>
<h2 id="limitation">LIMITATION</h2>
<p>All native Windows platforms emulating Linux are known to be
slow.</p>
<h2 id="bugs">BUGS</h2>
<p>Probably. Current known ones and interface for filing new
ones: https://testssl.sh/bugs/ .</p>
<h2 id="see-also">SEE ALSO</h2>
<p><code>ciphers</code>(1), <code>openssl</code>(1),
<code>s_client</code>(1), <code>x509</code>(1),
<code>verify</code>(1), <code>ocsp</code>(1),
<code>crl</code>(1), <code>bash</code>(1) and the websites
https://testssl.sh/ and https://github.com/testssl/testssl.sh/
.</p>
</body>
</html>