Finding and attribution
A small font with many avar breakpoints and a 65,535-row region-less VarData caused dense tuple allocation; the maintainer reports a 20 KB input reaching 2.8 GB peak memory.
Product: HarfBuzz. AI system: Anthropic OSS Scanner (model not identified for this discovery). Team / organization: Anthropic / HarfBuzz maintainers.
CVE identifiers: Not established in the cited evidence.
AI and human contribution (source account): The primary maintainer commit explicitly says “Discovered by Anthropic's OSS Scanner, as vulnerability ANT-2026-3HO8JDDV.” It separately lists Claude Opus 4.8 as a commit co-author; the source does not establish that model as the discovering system or describe its investigative steps.
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 maintainer source explicitly calls this a vulnerability and describes resource exhaustion from crafted font inputs in the subset path. It does not report exploitation in the wild.
Remediation status: fix documented. The maintainer commit documents a separate budget for avar2 tuple-memory deltas, rejecting the subset cleanly on exhaustion; it also reports the attack reproducer reduced from 2.8 GB to 21 MB. The commit does not establish an affected or fixed release boundary.
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 Nov 2026. AI attribution dated 12 Nov 2026. Sources checked 17 Nov 2026; these are separate dates, not a live installation check.
Recorded CVSS: Not available · no assessment established here. 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.
No source-backed CVSS assessment is established in this catalog.
Case-level evidence: denial of service.
Prerequisites / affected scope: A crafted font supplied to avar2 subsetting can drive excessive memory allocation; maintainer reports 2.8 GB peak memory from a 20 KB input..
Exploitation evidence: The cited maintainer source reports resource consumption from crafted test inputs; it does not report exploitation in the wild.. Impact source · checked 17 Nov 2026.
Documented fix: The maintainer commit documents a separate budget for avar2 tuple-memory deltas, rejecting the subset cleanly on exhaustion; it also reports the attack reproducer reduced from 2.8 GB to 21 MB. The commit does not establish an affected or fixed release boundary.