TLS vulnerability, protocol and certificate audit

Runs in: Deep TLS Security Audit

The TLS Security Audit runs a TLS test engine against one host and port of your site and sorts its findings into six sections: the engine's rating (grade and score), certificate, protocol support, named TLS vulnerabilities, cipher suites and HTTP security headers. The vulnerability section carries the engine's checks for Heartbleed, CCS injection, Ticketbleed, ROBOT, secure and client-initiated renegotiation, CRIME, BREACH, POODLE, TLS fallback SCSV, SWEET32, FREAK, DROWN, Logjam, BEAST, Lucky Thirteen, RC4 and WinShock. The audit runs on demand and in scheduled reports; it requires a login and the Pro plan.

How it tests

  • The audit connects to your host on the port you give (443 by default). Which TLS handshakes and probes are sent for each check is decided by the test engine; the audit passes no options that add or remove checks, apart from fast mode.
  • Fast mode: an on-demand run defaults to fast mode; a scheduled report defaults to the full (non-fast) run.
  • Before the engine starts, your host name is resolved. Private, loopback, link-local, IPv6 site-local (fec0::/10), multicast, reserved, unspecified and shared (100.64.0.0/10) addresses, and every other address that is not globally routable, are refused, and a refused host ends the run with that error; no test is sent. The engine connects only to the checked address; your host name is still sent as the TLS server name.
  • Each finding the engine returns is placed in one of the six sections. The Rating section is shown first: grade, then score, then each reason the engine gives for capping the grade. Engine output that belongs to none of the six sections is not shown.
  • Severity mapping: engine severities OK and INFO are shown as Passed; WARN, LOW and MEDIUM as Warning; HIGH, CRITICAL and NOT OK as Error; an entry with any other severity is not shown. Two exceptions in the Rating section: the score takes the status of the grade, and every grade cap reason or grade warning is shown as Warning.
  • Each entry carries a label (for example Heartbleed (CVE-2014-0160), CCS Injection (CVE-2014-0224), Ticketbleed (CVE-2016-9244), WinShock (CVE-2014-6321), POODLE (SSLv3), ROBOT Attack) and the engine's finding text, with the engine's CVE ids appended in brackets when the finding text does not already contain them.
  • A run is stopped after 600 s.
  • You can run the audit only against hosts under the domains of sites whose ownership your account has verified.

What you see

Situation Result
Full run completed section Rating: Overall Grade A+ to F (A grades Passed, B and C Warning, D to F Error), Final Score (0–100) with the grade's status, one Grade Cap Reason / Grade Warning (Warning) per reason the engine gives
Engine rates a vulnerability check HIGH, CRITICAL or NOT OK Error in section Vulnerabilities, e.g. Heartbleed (CVE-2014-0160) with the engine's finding text and CVE ids
Engine rates a check WARN, LOW or MEDIUM Warning with the engine's finding text
Engine rates a check OK or INFO Passed with the engine's finding text
Protocol, certificate, cipher or header entry same severity mapping, in sections Protocol Support, Certificate, Cipher Suites, HTTP Security
Engine missing error: The TLS audit engine is not installed on this server.
Engine wrote no output error: The TLS audit produced no output (exit <code>).
Output is not valid JSON error: Could not parse the TLS audit output: …
Run exceeds 600 s on demand, the run ends with a timeout; in scheduled reports, The TLS audit did not finish within 630 s

Limits

  • Verdicts are the engine's severities, mapped one to one. The audit applies no judgement of its own; LOW is shown as Warning, and an engine severity outside the mapped set is dropped.
  • Fast mode is the on-demand default. Which checks fast mode omits is defined by the engine.
  • These checks are reported, but they are not counted as coverage of a specific CVE: ROBOT and RC4 each span several CVE ids; client-initiated renegotiation, TLS fallback SCSV and the DROWN hint are configuration checks.
  • Only the first public address of your host is tested; further addresses of the same name are not.
  • A report threshold cannot reference an Error finding of this audit, and a newly vulnerable result raises no incident by itself.
  • The grade is the engine's own rating. When the engine returns no grade, the Rating section is not shown.

Vulnerabilities this check covers 14

Each advisory id links to its public record where one exists. This product uses data from the NVD API but is not endorsed or certified by the NVD.