Finding and attribution
The decoder processed every coded picture in an item or sequence sample even though it returns one image. A 16 kB file with 1,000 hidden HEVC pictures of 4096×4096 reportedly took 14 seconds.
Product: libheif. AI system: Anthropic OSS Scanner (model not specified for this finding). Team / organization: Anthropic and libheif maintainers.
CVE identifiers: Not established in the cited evidence.
AI and human contribution (source account): The item-level libheif advisory explicitly attributes discovery to Anthropic's OSS Scanner as ANT-2026-9W6QJITH. The release notes independently state that all listed security issues were found by the scanner. The Anthropic program sources establish model-based scanning, but no per-case model is identified.
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 primary advisory establishes a publicly disclosed security finding and direct Anthropic OSS Scanner attribution. No CVE had been assigned in the libheif release notes at the checked time; the distinct GHSA is used. No exploitation or zero-day status is asserted.
Remediation status: fix documented. Decoding now refuses more than eight coded pictures per decode by default. The v1.24.x limit can be configured; zero disables it. Fixed in libheif 1.23.7. The libheif v1.23.7 security release is dated 2026-10-26; 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 26 Oct 2026. AI attribution dated 26 Oct 2026. Sources checked 14 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 HEIF item or sequence sample causes a decode call to process hundreds or thousands of coded pictures; untrusted-file decoder workers can be occupied for seconds or minutes. Source says no memory corruption or confidentiality impact is known..
Exploitation evidence: The cited source does not establish exploitation in the wild.. Impact source · checked 16 Nov 2026.
Documented fix: Decoding now refuses more than eight coded pictures per decode by default. The v1.24.x limit can be configured; zero disables it. Fixed in libheif 1.23.7. The libheif v1.23.7 security release is dated 2026-10-26; deployed patch status is not checked.