CVE Index中文
L3 · ShartedClamAV · antivirus · DoS · Ciscoclamav-november-scanner-batch

ClamAV fixes ten scanner flaws in 1.5.5 and 1.4.7

ClamAV 1.5.5 and 1.4.7 patch ten flaws involving malformed scan inputs, memory errors, loops and resource-limit bypasses. Cisco Secure Endpoint Connectors require separate platform updates. Cisco reports no known malicious use.

Disclosed
Nov 18, 2026
Public disclosure date
Severity
CVSS 7.5 × 10
10 CVEs tracked here
Splash zone
ClamAV processing attacker-supplied scan inputs; some paths need phishing database entries or debug logging
ClamAV · antivirus · DoS · Cisco
In the wild
Not seen
Cisco PSIRT reports no known malicious use of the vulnerabilities.

What happened

ClamAV’s November 18 security release fixes ten independent vulnerabilities in 1.5.5 and the 1.4.7 backport. Exposure requires attacker-controlled content to reach the corresponding scanner, unpacker or logging path. The phishcheck issue additionally requires full-hash phishing database entries; Cisco describes the logging issue with debug logging enabled. A deployment affected by one bug need not be affected by all ten.

The upstream blog identifies historical introduction versions for each issue and describes the affected development range through 1.5.4, while explicitly identifying 1.4.7 as a fixed backport. Use the release-line-specific fixes above rather than treating the largest version in that historical range as the only upgrade target. These are engine fixes; the release announcement also warns that updated Docker images may not be immediately available.

The sources describe different impact detail. ClamAV documents stack/heap writes, a logging format-string problem that can disclose or corrupt process memory, scan loops and resource exhaustion. Cisco’s downstream advisory and CNA records describe denial of service and scan-limit bypasses in its products. In particular, the UPX report differs on read versus write, OLE2 differs on memory versus stack exhaustion, and XAR differs on file-size/looping versus extraction-storage limits. Preserve those product-specific descriptions; neither a memory-write label nor Windows scanner privileges establish a demonstrated remote code-execution chain here.

The displayed 7.5 scores are Cisco’s CVSS 3.1 values for the ten identifiers, not a uniform assessment of every deployment. Cisco rates Windows Connector impact High and Linux/Mac Medium, listing platform scores of 7.5 and 5.3 respectively. Cisco says Secure Endpoint Private Cloud itself is not impacted: its connector distribution is the concern. Its fixed-software table still has a Private Cloud release placeholder, so follow the Connector targets and content-update process instead of inventing a server patch version.

Cisco reports no known malicious use or public announcements of exploitation, and lists no workaround addressing the vulnerabilities. L3 reflects remotely supplied input reaching a mainstream scanning engine, with documented memory-safety and availability impact; it is not a report of victim compromise. No incident-specific attack indicators or victim count are published in these sources.

What’s in the bowl

Phishcheck URL length accounting

Crafted email can cause a stack-buffer write when full-hash phishing database entries are loaded. Upstream traces introduction to 0.94.

CHM cyclic directory traversal

Crafted CHM directory pointers can cause an infinite scanning loop. Upstream traces introduction to 0.100.0.

yC executable offset validation

Malformed yC-packed executables can trigger underflow and out-of-bounds heap reads. Upstream traces introduction to 0.96.

InstallShield worker loop

Malformed InstallShield file-table counts or offsets can pin a scan worker. Upstream traces introduction to 0.96.

OLE2 resource exhaustion

Upstream describes stack exhaustion from deeply nested OLE2 properties; Cisco describes memory exhaustion. Upstream traces introduction to 0.71.

CHM LZX match offsets

Malformed CHM LZX matches can cause out-of-bounds heap writes. Upstream traces introduction to 0.80.

Rust-to-C logging format handling

Upstream describes crash, process-memory disclosure or corruption from attacker-controlled logged text; Cisco describes a debug-logging DoS path. Introduced in 0.105.0.

UPX decompression and PE rebuild bounds

Upstream describes out-of-bounds heap writes while unpacking malformed executables; Cisco describes an out-of-bounds read/DoS. Introduced in 0.94.1.

OLE2 directory-chain scan-time bypass

Crafted OLE2 directory chains can consume excessive CPU beyond configured scan time. Upstream traces introduction to 0.98.1.

XAR extraction resource-limit bypass

Upstream describes temporary-storage exhaustion during extraction; Cisco describes MaxFileSize bypass and possible looping. Introduced in 0.98.

Fixed versions

ProductSafe fromNote
ClamAV 1.51.5.5November security fixes
ClamAV 1.41.4.7November security fixes backported
Cisco Secure Endpoint Connector for Linux1.29.3Cisco-listed first fixed release
Cisco Secure Endpoint Connector for Mac1.27.5Cisco-listed first fixed release
Cisco Secure Endpoint Connector for Windows8.6.4Cisco-listed first fixed release

Crime-scene timeline

  1. Nov 18, 16:00 UTC

    Cisco publishes its downstream advisory

    Ten independent ClamAV vulnerabilities affect Secure Endpoint Connectors; Cisco lists platform-specific fixes and no workaround.

  2. Nov 18, 16:30 UTC

    ClamAV publishes the upstream security release

    The project announces 1.5.5 and 1.4.7 with the same ten CVE fixes. Docker images may appear later than the announcement.

Who said what

We list each claim and who made it. Primary sources (the vendor, the researcher, the registry) are marked. Where reports disagree, we grade by the worst credible case.

  • ClamAV November security releasePrimaryNames all ten fixes, historical introductions, upstream 1.5.5/1.4.7 targets and container availability caveat; describes several memory-write and logging effects.
  • Cisco downstream ClamAV advisoryPrimaryDescribes DoS/limit-bypass impact, per-platform Connector fixes, Windows-versus-Linux/Mac severity, no workaround and no known malicious use.