HDF5 EUVD-2026-116008

Exploitation requires an attacker to supply a crafted HDF5 file and have a user or application open it via H5Fopen(); the file must contain an object-header message whose declared size passes the pre-header bounds check yet whose body begins up to four bytes further into the chunk (after the message header flags, and for version 2 object headers the optional creation index) so that it runs past the end of the chunk image, reaching a decode callback such as H5O__mtime_new_decode(). Limiting factors that reduce realistic risk: the vector is local (AV:L) and demands user interaction (UI:P - the victim must choose to open the malicious file); no special non-default configuration is required since any default H5Fopen() parse path is reachable; and impact is bounded to a crash or limited disclosure of process memory, with no code-execution or integrity impact established in the available sources.

All signals point to a genuinely low-priority issue. The supplied CVSS 4.0 base score is 1.9 (Low) with vector AV:L/AC:L/AT:N/PR:N/UI:P/VC:L/VI:N/VA:L/SC:N/SI:N/SA:N: the attack is local and requires passive user interaction (the victim must open a crafted file), and the only impacts are low confidentiality (limited process-memory disclosure) and low availability (a crash/DoS) - there is no integrity impact and no scope change to other systems. The vuln.today Decision label is 'monitor', the lowest urgency tier, and no CISA KEV listing or EPSS score was provided in the input, so there is no evidence of active or widespread exploitation. The one signal above baseline is the threat/exploit-maturity metric E:P (Proof-of-Concept) in the CVSS 4.0 vector, which is consistent with the upstream fix shipping a regression test that reproduces the over-read; this indicates the bug is demonstrable, not that it is being exploited. In short, this is a low-CVSS, low-real-risk file-parsing defect rather than an emergency. Note that no EPSS value and no KEV status were supplied - their absence is expected for a newly-fixed low-severity library bug and is not treated as exploitation evidence.

An attacker crafts a malicious HDF5 file that declares an object-header message size which passes the initial bounds check but whose body extends past the end of the chunk image, then delivers it to a victim (shared dataset, download, or email attachment). When the victim's application calls H5Fopen() on the file, a decode callback such as H5O__mtime_new_decode() reads beyond the heap allocation, crashing the process or leaking a small amount of adjacent process memory. No standalone weaponized exploit is identified; the upstream fix includes a regression test that reproduces the over-read under a sanitizer.