SSLv3 Padding Oracle On Downgraded Legacy Encryption Vulnerabi

Tenable plugin 78479 is flagging SSLv3 support and mapping it to POODLE / CVE-2014-3566. The underlying flaw is in the SSL 3.0 protocol design: if a server still negotiates SSLv3 with CBC ciphers, an active man-in-the-middle can force a downgrade and gradually recover small pieces of plaintext such as session cookies. This is not tied to one app build; it affects any service that still allows SSLv3, and older OpenSSL trains only added downgrade signaling (TLS_FALLBACK_SCSV) rather than making SSLv3 itself safe.

The original medium-era framing overstated what matters in enterprise prioritization. In real environments, this attack requires on-path network position, downgrade success, and usually a real user session to harvest anything useful. Modern browsers and most managed server baselines already killed SSLv3 years ago, so what Tenable finds now is usually legacy infrastructure debt, not a broadly weaponizable internet edge bug. Keep it on the cleanup list, but don't let a protocol hygiene finding crowd out remotely exploitable initial-access bugs.

Why this verdict

• Requires on-path interception — this is not unauthenticated remote exploitation from anywhere on the internet; the attacker must already control or influence the victim's network path.
• Requires downgrade behavior — both the client and server have to cooperate badly enough for a fallback to SSLv3, and modern browsers/libraries usually do not.
• Impact is narrow — the practical outcome is usually cookie or session disclosure, not code execution or host compromise.
• Legacy exposure only — if you still have this finding, it is usually on old appliances, embedded services, or forgotten admin interfaces rather than mainstream front-door applications.
• Scanner finds configuration debt, not a live exploit path — Tenable is correctly flagging unsafe protocol support, but the presence of SSLv3 alone does not imply a remotely exploitable attack chain in a modern estate.

Why not higher?

To score higher, this would need to behave like a reliable initial-access or broad post-auth break. It does not: the attack chain is gated by MITM position, downgrade success, and usually a real user session. Those prerequisites compound, and each one cuts the reachable population hard.

Why not lower?

I would not mark it IGNORE because SSLv3 should not still be present on managed enterprise services. If the scanner sees it, you have a real cryptographic hygiene failure that can matter on hostile internal networks, guest Wi-Fi, or legacy client paths. LOW is the right bucket for 'fix it, but do not let it jump the queue.'

1. Disable SSLv3 everywhere — Turn off SSLv3 on servers, reverse proxies, load balancers, Java containers, mail gateways, and any SSL/TLS-terminating middleware. For a LOW verdict there is no fixed mitigation SLA, so treat this as backlog hygiene and remove it during the normal hardening cycle rather than as an emergency change.
2. Terminate TLS upstream on modern infrastructure — If the backend app or appliance is old, move TLS termination to a current reverse proxy or load balancer that only allows TLS 1.2/1.3. This is often the fastest way to kill exposure without waiting on a vendor firmware refresh; for LOW, schedule it as normal engineering work.
3. Block downgrade paths — Where legacy compatibility constraints exist, ensure current libraries and endpoints support TLS_FALLBACK_SCSV and enforce minimum protocol floors. This does not replace disabling SSLv3, but it reduces exploitability while you retire the weak protocol.
4. Hunt for legacy listeners — Inventory non-443 services, admin ports, embedded HTTPS stacks, IPMI/iDRAC-era interfaces, and appliance management planes. Most surviving SSLv3 findings live off the main web path, so close the long tail instead of assuming the public VIPs are the whole story.

Run this from an auditor workstation or jump host that can reach the target service over the network. Invoke it as ./check_poodle.sh example.com 443; no privileges are needed beyond network access to the service, but your local openssl binary must still support the -ssl3 test mode.

#!/usr/bin/env bash
# check_poodle.sh
# Detects whether a remote service still negotiates SSLv3, which implies POODLE exposure.
# Usage: ./check_poodle.sh <host> [port]
# Exit codes: 0=PATCHED, 1=VULNERABLE, 2=UNKNOWN, 3=usage error

set -u

HOST="${1:-}"
PORT="${2:-443}"
TIMEOUT_BIN=""

if [[ -z "$HOST" ]]; then
  echo "Usage: $0 <host> [port]"
  exit 3
fi

if command -v timeout >/dev/null 2>&1; then
  TIMEOUT_BIN="timeout 10"
fi

if ! command -v openssl >/dev/null 2>&1; then
  echo "UNKNOWN - openssl not installed"
  exit 2
fi

# Check whether this OpenSSL build can even attempt SSLv3.
if ! openssl s_client -help 2>&1 | grep -q -- '-ssl3'; then
  echo "UNKNOWN - local openssl does not support '-ssl3'; test from a host with legacy protocol probing support"
  exit 2
fi

OUTPUT=$(bash -c "$TIMEOUT_BIN openssl s_client -connect ${HOST}:${PORT} -ssl3 -servername ${HOST} </dev/null" 2>&1)
RC=$?

# Timeout or connectivity issue
if [[ $RC -eq 124 || $RC -eq 137 ]]; then
  echo "UNKNOWN - connection timed out probing ${HOST}:${PORT}"
  exit 2
fi

if echo "$OUTPUT" | grep -Eqi 'handshake failure|wrong version number|unsupported protocol|no protocols available|alert protocol version|sslv3 alert handshake failure'; then
  echo "PATCHED - SSLv3 negotiation failed on ${HOST}:${PORT}"
  exit 0
fi

if echo "$OUTPUT" | grep -Eqi 'Protocol *: *SSLv3|New,.*Cipher is|SSL-Session:'; then
  echo "VULNERABLE - ${HOST}:${PORT} accepted SSLv3"
  exit 1
fi

if echo "$OUTPUT" | grep -Eqi 'connect:errno=|Connection refused|Name or service not known|Temporary failure in name resolution|getaddrinfo'; then
  echo "UNKNOWN - could not reach ${HOST}:${PORT}"
  exit 2
fi

echo "UNKNOWN - inconclusive result probing ${HOST}:${PORT}"
exit 2

Sources

1. Tenable Nessus Plugin 78479
2. NVD CVE-2014-3566
3. Google disclosure: This POODLE bites
4. Mozilla: The POODLE Attack and the End of SSL 3.0
5. Red Hat advisory and mitigation guidance
6. CISA alert on SSL 3.0 / POODLE
7. Public PoC repository
8. Qualys SSL Pulse