CVE-2026-55213
h2o is an HTTP server with support for HTTP/1.x, HTTP/2 and HTTP/3. Prior to commit edd7a120bfc4af11ac0cbebce2a43cc1f93f9af1, when h2o proce
CVSS
7.5
High
EPSS
0.3%
p27
KEV
—
Exploit Today
8
0-100
Published: Jul 10, 2026 · Last modified: Jul 13, 2026 · CWE-789
0.3%EPSS · 30 days0.3%
2026-07-112026-07-21
h2o is an HTTP server with support for HTTP/1.x, HTTP/2 and HTTP/3. Prior to commit edd7a120bfc4af11ac0cbebce2a43cc1f93f9af1, when h2o processes a QPACK instruction sent from the peer over HTTP/3, lib/http3/qpack.c might allocate an on-stack buffer as large as approximately 800 KB by calling alloca, which exceeds the default pthread stack size used by musl libc and causes the h2o server to crash with a segmentation fault while touching the guard page. This issue is fixed in commit edd7a120bfc4af11ac0cbebce2a43cc1f93f9af1.
CVECVSSEPSSKEVRExploitTitleMod.
CVE-2026-653157.5 HIG—
——0Ollama (HEAD f0078ae) contains an uncontrolled memory allocation vulnerability in the GGUF metadata parser that allows remote attackers to crash the server by supplying a crafted GGUF file with attacker-controlled length and count fields in string lengths, tensor dimension counts, and metadata array counts that are used as allocation sizes without validation against remaining file size. Attackers can upload a sub-1KB crafted GGUF file via the blob upload and model create or pull API endpoints to trigger unrecoverable Go runtime out-of-memory fatal errors or makeslice panics that bypass recovery middleware and crash the entire server process.12hCVE-2026-476677.5 HIG—
——0CImg Library is a C++ library for image processing. Prior to version 4.0.0 in `_load_analyze()`, the header_size field is read as an `unsigned int` from the first 4 bytes of an Analyze/NIfTI file and passed directly to `new unsigned char[header_size]` without being bounded against the actual file size. A value up to ~4 GB is accepted. If the subsequent `fread` returns `short` as it will for any malformed file), the function throws a `CImgIOException` and the allocated buffer is never freed. A 6-byte crafted file is sufficient to trigger an allocation of ~1.3 GB per call, with the full allocation leaked on every error path. The issue is reachable via `load_analyze()` and the generic `load()` when the file extension is .hdr, .img, or .nii. Version 4.0.0 fixes the issue.14hCVE-2026-598446.5 MED—
——0A flaw was found in libssh. A remote authenticated client can issue SSH_FXP_READ requests with an arbitrarily large length, causing a libssh SFTP server to allocate excessive memory and potentially exhaust it through repeated requests.15hCVE-2025-71395—14.4%
——4SurrealDB versions before 2.2.2 contain a memory exhaustion vulnerability in the string::replace function that fails to restrict resulting string length when using regex patterns. An authenticated attacker can craft a malicious query to exhaust server memory through unbounded string allocations, causing denial of service.15hCVE-2026-444537.5 HIG19.9%
——6h2o is an HTTP server with support for HTTP/1.x, HTTP/2 and HTTP/3. Prior to commit 6b5370d, h2o is vulnerable to a Denial of Service attack when calling alloca under certain conditions. When serving static files, h2o builds the file path on stack, by calling alloca. The maximum size of the memory allocated using alloca can be as huge as ~600KB, which exceeds the default pthread stack size used by musl libc (128KB). If the amount of memory allocated by alloca exceeds the stack size, the h2o server crashes with a segmentation fault, while it tries to touch the guard page. This issue has been fixed by commit 6b5370d.5dCVE-2026-55407—43.3%
——13Buffa is a pure-Rust Protocol Buffers implementation with first-class protobuf editions support. Prior to 0.8.0, the decode_unknown_field function in buffa's protobuf decoder allocated heap memory in proportion to untrusted input (unknown fields in the serialized protobuf) without enforcing an allocation budget, affecting any message decoded from untrusted input using code generated with preserve_unknown_fields=true (the default); a small, well-formed payload of nested unknown fields inside a StartGroup could trigger roughly 22x memory amplification (for example a 64 MiB input forcing about 1.4 GB of heap allocation), and length-delimited unknown fields could be sized arbitrarily, so an unauthenticated attacker could crash a process through memory exhaustion because the top-level message size cap did not account for in-decode amplification. This issue is fixed in version 0.8.0.6d