Finding and attribution
Wireshark's official advisory describes a crash when opening a malformed TTX Logger capture and says it was discovered by Claude and Ada Logics; it reports no known exploits. Anthropic's item-level CVD report maps the issue to the TTL segmented-message reader: a file-controlled small allocation is followed by fixed-offset reads and an attacker-controlled timestamp write, producing an 8-byte out-of-bounds read and write. Anthropic notes static analysis rather than dynamic reproduction in its report; the external security research firm later classified the finding as a true positive. The impact is through opening a crafted file, not passive network traffic.
Product: Wireshark TTX Logger (.ttl) file parser. AI system: Claude (Anthropic; Anthropic CVD identifies the model as Claude Mythos Preview). Team / organization: Anthropic, Ada Logics, and Wireshark.
CVE identifiers: CVE-2026-19694.
AI and human contribution (source account): Wireshark’s public Aug 12 vendor advisory credits discovery to Claude and Ada Logics. Anthropic’s item-level case ANT-2026-EKGJXN5A later names Claude Mythos Preview and says Ada Logics performed triage/disclosure. The Anthropic report initially describes static analysis rather than dynamic reproduction; the campaign page records Ada Logics’ true-positive verdict. The vendor-linked Wireshark work item #21389 resolves the name/function mapping: it titles the issue as the TTL (TTX Logger) capture reader, identifies `wiretap/ttl.c` and `ttl_fix_segmented_message_entry_timestamp()`, and explicitly says “found by Anthropic using Claude”; it attaches ANT-2026-EKGJXN5A and records Ada Logics manual review.
Human involvement is only described where the source account establishes it; unknown involvement does not mean none. Vendor confirmation of a flaw does not confirm every claim about AI contribution.
Discovery context: The vendor advisory establishes the published issue, Claude credit, and fixed release. Anthropic's later public report supplies the specific Mythos model attribution and details the memory-safety behavior, while explicitly describing its initial report as static analysis; Ada Logics subsequently marked it true positive. No exploit activity is asserted; Wireshark says it was unaware of exploits.
Remediation status: fix documented. Wireshark lists version 4.6.8 as fixed on Aug 12, 2026. Upgrade to 4.6.8 or later; deployed patch status is not checked.
A source-documented fix does not establish installed patches or current safety. Source checking does not independently reproduce the finding or establish observed exploitation.
Disclosed 12 Aug 2026. AI attribution dated 12 Aug 2026. Sources checked 14 Nov 2026; these are separate dates, not a live installation check.
Recorded CVSS: Medium · highest recorded score 4.7. Documented technical effects: denial of service.
Severity, consequences and exploitation evidence
Case grouping uses the highest recorded CVSS score, with its version-specific severity band. It is descriptive, not a combined or editorial risk index. All recorded assessments remain below; scores from different versions and sources may disagree. CVSS specifications define version-specific scoring and severity bands.
- CVE-2026-19694: CVSS 3.1 · 4.7 (Medium) ·
CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:N/I:N/A:H· Assessment scenarios: GENERAL. GitLab (CNA; CVE Program record) · checked 16 Nov 2026.
CVE-2026-19694: denial of service.
Prerequisites / affected scope: Opening a malformed TTX Logger capture crashes affected Wireshark versions..
Exploitation evidence: The cited source does not establish exploitation in the wild.. Impact source · checked 16 Nov 2026.
Documented fix: Wireshark lists version 4.6.8 as fixed on Aug 12, 2026. Upgrade to 4.6.8 or later; deployed patch status is not checked.