Transport Layer Security

tenable:62565 is not telling you the server is actively exploitable on its own; it is telling you the service appears to meet one of CRIME's old prerequisites: negotiated TLS compression or legacy SPDY earlier than version 4. The underlying issue is a compression side channel that can leak secrets such as HTTP session cookies when compressed data is encrypted and an attacker can repeatedly influence requests and measure ciphertext lengths. In practice that means TLS 1.2-and-earlier stacks with compression enabled, or very old SPDY deployments, both of which are now mostly legacy edge cases rather than broad enterprise defaults.

The vendor's LOW rating is basically right. The attack needs a man-in-the-middle position, a victim browser you can coerce into sending many chosen requests, and a target that still negotiates obsolete compression behavior; modern defaults kill this chain in most environments. This is worth cleaning up if the finding is real, but it should not outrank remotely exploitable server-side RCE, auth bypass, or modern edge-service bugs.

Why this verdict

• Needs MITM first: this is not an unauthenticated internet-to-server exploit chain. Requiring on-path traffic control is major downward pressure from the vendor baseline.
• Feature retirement crushes exposure: real exploitation needs negotiated TLS compression or old SPDY. OpenSSL disabled compression by default in 1.1.0, and Chrome removed SPDY/NPN in Chrome 51, so only legacy stacks remain.
• Blast radius is narrow: successful exploitation usually steals one user's session cookie. There is no server-side code execution, no wormability, and no straightforward path to host takeover.

Why not higher?

There is no credible case for pushing this above LOW in a modern enterprise unless you have a very specific legacy internet-facing application with confirmed TLS compression and high-value browser sessions over hostile networks. Every important prerequisite narrows the attack population: attacker position, victim interaction, protocol behavior, and session stability.

Why not lower?

It is still a real confidentiality weakness if the prerequisite is truly present. A hostile on-path attacker can turn it into session hijack against exposed user workflows, so completely ignoring a confirmed internet-facing instance would be too dismissive.

1. Disable TLS compression — Turn off TLS-level compression on the actual terminator that answered the scan: load balancer, reverse proxy, web server, VPN gateway, or embedded appliance. For a LOW verdict there is no formal SLA, so fold this into the next routine TLS hardening window rather than an emergency change.
2. Retire SPDY and NPN — If any legacy edge device still advertises SPDY or old NPN behavior, remove it and standardize on HTTP/2 or HTTP/3. In 2026 this is backlog hygiene work; schedule it in the next normal platform cleanup cycle.
3. Harden session cookies — Use SameSite, short session lifetimes, rotation after login, and reauthentication for sensitive actions to reduce the usefulness of any stolen cookie. This does not fix the root cause, but it cuts the practical value of a successful oracle attack during the next application security sprint.
4. Validate the finding at the edge — Confirm which device actually negotiates TLS for the reported port before assigning work. Many environments terminate TLS on ADCs or CDN edges, so the right owner is often network/platform engineering rather than the app team.

Run this from an auditor workstation that can reach the target TCP port; no root is required. Invoke it as ./check_crime_tls.sh example.com:443 or ./check_crime_tls.sh 10.20.30.40:8443. It verifies the TLS-compression branch of the CRIME prerequisite using openssl; if your local OpenSSL lacks compression support, it returns UNKNOWN rather than guessing.

#!/usr/bin/env bash
# check_crime_tls.sh
# Verify whether a target negotiates TLS compression, a prerequisite for CRIME.
# Output: VULNERABLE / PATCHED / UNKNOWN
# Exit codes: 1=VULNERABLE, 0=PATCHED, 2=UNKNOWN

set -u

TARGET="${1:-}"
if [[ -z "$TARGET" ]]; then
  echo "UNKNOWN"
  exit 2
fi

if ! command -v openssl >/dev/null 2>&1; then
  echo "UNKNOWN"
  exit 2
fi

HOST="$TARGET"
PORT="443"
if [[ "$TARGET" == *:* ]]; then
  HOST="${TARGET%%:*}"
  PORT="${TARGET##*:}"
fi

# Use -comp to request compression if the local OpenSSL supports it.
# Lower security level only for the handshake probe; this does not change the target.
CMD=(openssl s_client -connect "${HOST}:${PORT}" -servername "$HOST" -tls1_2 -comp -cipher 'DEFAULT:@SECLEVEL=1')

OUT=$(timeout 15 bash -c "printf '' | \"${CMD[0]}\" ${CMD[*]:1}" 2>&1)
RC=$?

if [[ $RC -ne 0 && $RC -ne 1 ]]; then
  echo "UNKNOWN"
  exit 2
fi

# If local OpenSSL was built without compression capability, -comp may be unsupported
# or the handshake output may not include a Compression line.
COMP_LINE=$(printf '%s\n' "$OUT" | awk '/Compression:/ {print; found=1} END{if(!found) print ""}')

if [[ -z "$COMP_LINE" ]]; then
  echo "UNKNOWN"
  exit 2
fi

if printf '%s\n' "$COMP_LINE" | grep -Eq 'Compression:\s*NONE'; then
  echo "PATCHED"
  exit 0
fi

if printf '%s\n' "$COMP_LINE" | grep -Eq 'Compression:\s*(zlib|ZLIB|DEFLATE|[A-Za-z0-9_-]+)'; then
  echo "VULNERABLE"
  exit 1
fi

echo "UNKNOWN"
exit 2

Sources

1. Tenable Nessus Plugin 62565
2. NVD CVE-2012-4929
3. Qualys CRIME analysis
4. OpenSSL `SSL_CONF_cmd` documentation
5. Chromium Blog: Transitioning from SPDY to HTTP/2
6. RFC 3749: TLS Compression Methods
7. CISA Known Exploited Vulnerabilities Catalog
8. Snyk vulnerability page with EPSS reference for CVE-2012-4929