Newly released CVEs across every platform — sleek to read, verbose on demand, and served raw as markdown for AI and agent ingestion. Severity reads as depth: the deeper the contact, the graver the threat.
Depth = severity + exploitation
RUSTSEC-2026-0420Medium· 6.4wasip3 UDP `send` can deliver a datagram to an address only allowed for receiving
CVE-2026-58494Medium· 6.5Wasmtime: WASI hard links bypass wasmtime-wasi's FilePerms for destination
CVE-2026-118688High· 7.9Filesystem sandbox escape when renaming symlinks with trailing slashes
RUSTSEC-2026-0352NoneFilesystem sandbox escape when renaming symlinks with trailing slashes
CVE-2026-115386Low· 2.6fd_readdir copies uninitialized struct padding into guest memory
RUSTSEC-2026-0324Medium· 6.2Guest can panic host through filesystem timestamp before the epoch on wasip3
RUSTSEC-2026-0323Nonefd_readdir copies uninitialized struct padding into guest memory
RUSTSEC-2026-0322NoneExcessive allocated memory on the host when guests don't have stdio
RUSTSEC-2026-0321Medium· 4.0WASI preview 0 implementation of `poll_oneoff` circumvents fuel consumption
CVE-2026-115387Medium· 6.2Guest can panic host through filesystem datetime overflow
RUSTSEC-2026-0314Medium· 6.2Guest can panic host through filesystem datetime overflow
CVE-2026-54786LowLeak in WASIp1 `fd_renumber` implementation
CVE-2026-47261High· 7.5⚖ disputedWASI path_open(TRUNCATE) bypasses `FilePerms::WRITE` host restriction
The beta is over. VulnSea now runs at vulnsea.com, every old beta link redirects to the same page, and your account, API keys and watchlists came with it. Plus a round of hardening, faster pages, and new request limits agents should know about.
A summary of everything that shipped over the last two weeks — the whole corpus is open, agents get change feeds, alias resolution and EPSS movers, and the data now includes CVE.org, vendor CSAF, aggregated exploits and per-source scores.
A step-by-step guide to plugging VulnSea into automated and agentic workflows — poll the delta, triage without burning tokens, match an SBOM, and let an MCP-native model do the reasoning.