← Back to Feed CACHED · 2026-08-26 06:28:21 · CACHE_KEY CVE-2026-63076
CVE-2026-63076 · CWE-476 · Disclosed 2026-08-25

Issue summary: OpenSSL CMP password based protection verification only checks whether the protectionAlg…

ASSESSED — NOISGATE V0.5
Vendor
Reassessed
Verdict:
Do you agree?
01 · The Real Story

It is like a fire alarm wired only to a storage closet nobody visits

CVE-2026-63076 is an invalid pointer dereference in OpenSSL's Certificate Management Protocol (CMP) implementation. When verifying password-based MAC (PBM) protection on incoming CMP messages, OpenSSL calls X509_ALGOR_get0() to read the protectionAlg parameter but only checks whether the returned pointer is non-NULL — it never validates the ASN.1 type before casting to PBMParameter. A remote attacker can send a crafted CMP message with a mistyped protectionAlg field, causing a crash. Affected versions span OpenSSL 3.0.0–3.0.21, 3.4.0–3.4.6, 3.5.0–3.5.7, 3.6.0–3.6.3, and 4.0.0–4.0.1. Fixed in 3.0.22, 3.4.7, 3.5.8, 3.6.4, and 4.0.2.

The NVD-assigned CVSS 7.5 (HIGH) reflects the vector AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H, which is technically accurate for a network-reachable unauthenticated crash. However, OpenSSL's own project rates this Moderate, and that is far closer to reality. CMP is a specialized certificate enrollment protocol that applications must *explicitly enable* — the vast majority of OpenSSL consumers (web servers, TLS clients, VPN endpoints) never touch CMP code at all. The 7.5 score dramatically overstates risk for any organization that does not operate a CMP-enabled PKI enrollment service.

"CMP is a niche PKI protocol most shops never enable; this is a DoS-only crash in a rarely used feature."
02 · The Attack Path

3 steps from start to impact.

STEP 01

Identify a CMP endpoint

The attacker must locate a service that has explicitly enabled OpenSSL's CMP server functionality (or act as a malicious CMP server that a client connects to). CMP endpoints listen on HTTP and are used for certificate enrollment per RFC 4210. They are not exposed by default and must be purpose-built.
Conditions required:
  • Target runs an application using OpenSSL's CMP API
  • CMP endpoint is network-reachable to the attacker
Where this breaks in practice:
  • CMP is an opt-in feature; fewer than 1% of OpenSSL deployments use it
  • CMP servers are typically internal PKI infrastructure, not internet-facing
  • No Shodan/Censys signature widely indexes CMP endpoints
STEP 02

Craft a malicious CMP message

The attacker constructs a CMP PKIMessage with a protectionAlg field containing a valid algorithm OID but an ASN.1 parameter of the wrong type (e.g., an OCTET STRING instead of a PBMParameter SEQUENCE). The message must claim PBM-based protection so the server enters the vulnerable code path.
Conditions required:
  • Knowledge of CMP protocol framing (RFC 4210 / RFC 4211)
  • Ability to send raw CMP requests to the endpoint
Where this breaks in practice:
  • No public proof-of-concept exploit exists as of 2026-08-26
  • Requires niche protocol knowledge; off-the-shelf tools do not generate this payload
STEP 03

Trigger crash via invalid pointer dereference

When the CMP server processes the crafted message, X509_ALGOR_get0() returns a non-NULL pointer to the wrong ASN.1 structure. The code treats this as a PBMParameter and dereferences fields at incorrect offsets, causing a NULL or invalid pointer dereference and an immediate process crash. Impact is denial of service only — no memory disclosure, no controlled write, no code execution path.
Conditions required:
  • The server uses PBM-based message protection (password-based MAC)
Where this breaks in practice:
  • Process restart restores service immediately
  • Watchdog/systemd will auto-restart the crashed process in most deployments
  • Impact ceiling is availability loss of a single CMP endpoint
Detection/coverage: Application crash logs will show SIGSEGV in the CMP verification code path. Core dumps reference ossl_cmp_msg_check_update or related CMP functions.
03 · Intelligence Metadata

The supporting signals.

In-the-wild exploitationNone observed. Not listed in CISA KEV. No reports of active exploitation as of 2026-08-26.
Proof-of-conceptNo public PoC. No exploit code found on GitHub, Exploit-DB, or security researcher disclosures.
EPSSNot yet scored (disclosed 2026-08-25). Expected to be very low given niche attack surface.
KEV statusNot listed. No CISA KEV entry.
CVSS vectorCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H — 7.5 HIGH. Network-reachable, no auth, availability-only impact. OpenSSL project rates it Moderate.
Affected versionsOpenSSL 3.0.0–3.0.21, 3.4.0–3.4.6, 3.5.0–3.5.7, 3.6.0–3.6.3, 4.0.0–4.0.1
Fixed versions3.0.22, 3.4.7, 3.5.8, 3.6.4, 4.0.2
FIPS impactNone. CMP is outside the FIPS module boundary.
Scanning/exposureNo Shodan/Censys/GreyNoise signatures for CMP endpoints. CMP services are rarely internet-facing; exposure population is negligible.
Disclosure2026-08-25 via OpenSSL security advisory. Coordinated disclosure.
04 · The Call

noisgate verdict.

Final Verdict
DOWNGRADED to MEDIUM (4.0/10)

The single most decisive factor is the vanishingly small exposure population: CMP is an opt-in PKI enrollment protocol that fewer than 1% of OpenSSL deployments enable, and the impact ceiling is a recoverable process crash (DoS) with no path to code execution or data compromise.

HIGH Vulnerability mechanism and impact ceiling (DoS only, no RCE)
HIGH Affected/fixed version ranges (confirmed by OpenSSL advisory)
MEDIUM Exposure population estimate (<1% CMP usage across OpenSSL installs)

Why this verdict

  • Niche attack surface: CMP must be explicitly enabled by the application developer; standard TLS server/client usage of OpenSSL never invokes this code. This alone eliminates >99% of OpenSSL deployments from scope.
  • DoS-only impact ceiling: The vulnerability causes a process crash via invalid pointer dereference. There is no memory disclosure, no controlled write, and no viable path to remote code execution. A process restart fully restores service.
  • No exploitation leverage: No public PoC, no KEV listing, no observed campaigns. The protocol knowledge required (RFC 4210 CMP framing with ASN.1 type manipulation) is niche.
  • Role multiplier: CMP servers are PKI/CA infrastructure (high-value role). If the chain succeeds, the blast radius is *availability loss of certificate enrollment* — not domain takeover, not fleet compromise, not data exfiltration. A crashed CMP endpoint auto-restarts and issues no certificates during downtime (seconds to minutes). This limits the high-value-role outcome to temporary availability disruption, which does not meet the floor criteria of 'domain takeover / fleet compromise / mass data egress / supply-chain pivot.'
  • OpenSSL's own assessment agrees: The OpenSSL project rates this Moderate, not High, explicitly because of the limited attack surface.

Why not higher?

Upgrading to HIGH would require either active exploitation, a broader attack surface, or an impact beyond DoS. None of these conditions exist. CMP is not a default-enabled feature, no PoC is public, and the impact is a recoverable crash with no confidentiality or integrity consequences. Even in PKI environments, the blast radius is temporary enrollment downtime, not compromise of issued certificates or CA key material.

Why not lower?

Dropping to LOW would undercount the risk for the small number of organizations that *do* run CMP-enabled PKI enrollment services. In those environments, the vulnerability is unauthenticated and network-reachable, and a sustained attack could repeatedly crash the enrollment service. The breadth of affected OpenSSL versions (3.0 through 4.0) also means any CMP deployment is likely vulnerable.

05 · Compensating Control

What to do — in priority order.

  1. Restrict network access to CMP endpoints — Place CMP enrollment services behind firewall rules or a reverse proxy that limits source IPs to authorized RAs and enrollment clients. This eliminates unauthenticated internet-sourced attacks. Deploy within the noisgate remediation SLA of 365 days for MEDIUM, or sooner if your CMP service is externally reachable.
  2. Enable process watchdog/auto-restart — Ensure CMP server processes are managed by systemd, supervisord, or equivalent with automatic restart on crash. This limits DoS window to seconds per crash attempt. Most production deployments already have this.
  3. Rate-limit CMP HTTP requests — Apply rate limiting at the reverse proxy or load balancer layer to throttle repeated crash attempts and make sustained DoS impractical.
  4. Patch OpenSSL to fixed version — Upgrade to OpenSSL 3.0.22, 3.4.7, 3.5.8, 3.6.4, or 4.0.2 depending on your branch. This is the definitive fix. Target the noisgate remediation SLA of 365 days for MEDIUM severity.
What doesn't work
  • WAF rules — CMP uses binary ASN.1/DER encoding over HTTP, not text-based payloads. Standard WAF signatures cannot inspect or filter malformed CMP messages.
  • TLS client certificate authentication on the CMP endpoint — The vulnerability is in PBM (password-based MAC) verification, which occurs at the CMP protocol layer *after* TLS termination. mTLS does add friction but does not prevent a client with a valid cert from sending a malformed CMP message.
06 · Verification

Crowdsourced verification payload.

Run this script on each host where OpenSSL is installed. It checks the installed OpenSSL version against known-vulnerable ranges. Execute as any user: bash check_cve_2026_63076.sh. No elevated privileges required.

noisgate-verify.sh
BASHREAD-ONLYSAFE
#!/bin/bash\n# CVE-2026-63076 checker — Invalid Pointer Dereference in OpenSSL CMP\n# Run on target host. No privileges required.\n\nVERSION=$(openssl version 2>/dev/null | awk '{print $2}')\n\nif [ -z \"$VERSION\" ]; then\n  echo \"UNKNOWN — openssl binary not found in PATH\"\n  exit 2\nfi\n\necho \"Detected OpenSSL version: $VERSION\"\n\n# Parse major.minor.patch\nIFS='.' read -r MAJOR MINOR PATCH <<< \"$(echo \"$VERSION\" | sed 's/[^0-9.]//g')\"\n\nvulnerable=0\n\nif [ \"$MAJOR\" -eq 4 ] && [ \"$MINOR\" -eq 0 ]; then\n  [ \"$PATCH\" -lt 2 ] && vulnerable=1\nelif [ \"$MAJOR\" -eq 3 ] && [ \"$MINOR\" -eq 6 ]; then\n  [ \"$PATCH\" -lt 4 ] && vulnerable=1\nelif [ \"$MAJOR\" -eq 3 ] && [ \"$MINOR\" -eq 5 ]; then\n  [ \"$PATCH\" -lt 8 ] && vulnerable=1\nelif [ \"$MAJOR\" -eq 3 ] && [ \"$MINOR\" -eq 4 ]; then\n  [ \"$PATCH\" -lt 7 ] && vulnerable=1\nelif [ \"$MAJOR\" -eq 3 ] && [ \"$MINOR\" -eq 0 ]; then\n  [ \"$PATCH\" -lt 22 ] && vulnerable=1\nfi\n\nif [ \"$vulnerable\" -eq 1 ]; then\n  echo \"VULNERABLE — OpenSSL $VERSION is in the affected range for CVE-2026-63076\"\n  echo \"Note: Only exploitable if application explicitly uses CMP (Certificate Management Protocol)\"\n  exit 1\nelse\n  echo \"PATCHED — OpenSSL $VERSION is not affected or is a fixed release\"\n  exit 0\nfi
07 · Bottom Line

If you remember one thing.

TL;DR
For most enterprises, this CVE is a non-event — if you do not run CMP-enabled certificate enrollment services, the vulnerable code is never reached. First, determine if any application in your fleet actually uses OpenSSL's CMP API (search for OSSL_CMP_ symbols or openssl cmp CLI usage). If none do, document the rationale and move on. If you do operate CMP services, there is no noisgate mitigation SLA for MEDIUM — go straight to the 365-day noisgate remediation SLA by patching OpenSSL to the fixed version for your branch (3.0.22, 3.4.7, 3.5.8, 3.6.4, or 4.0.2). In the interim, restrict network access to CMP endpoints to authorized sources only and ensure process auto-restart is configured. This does not warrant emergency patching or change-freeze exceptions.

Sources

  1. OpenSSL Vulnerabilities Page (3.6 series)
  2. OpenSSL Vulnerabilities Index
  3. OpenSSL 3.0 Series Release Notes
  4. OpenSSL 3.6 Series Release Notes
  5. OpenSSL GitHub Releases
  6. Keyfactor CMP Operations Guide
  7. EJBCA Blog — CMP with OpenSSL
Peer Review

What defenders are saying.

Submit a review attribution: handle + country only
0 flags selected · stored anonymously
Validation Results

Crowdsourced verification outputs.

Results submitted by users who ran the verification payload against their environment.