Zebra CVE-2026-104437

Exploitation requires crafting a V5+ (post-NU5, ZIP-244) transparent transaction in which a SIGHASH_SINGLE (or SIGHASH_SINGLE|ANYONECANPAY) input has no corresponding output - i.e., fewer outputs than inputs at the signed index - which Zebra before 4.4.0 wrongly accepts instead of failing. No authentication is required (PR:N): any network peer can broadcast the transaction into Zebra's mempool. The critical LIMITING factor (reflected as AT:P in the CVSS 4.0 vector) is that the consensus-split impact only occurs if a miner running Zebra selects the malformed transaction via getblocktemplate and mines it into a block, which zcashd then rejects; the transaction alone against a non-mining Zebra node causes no block-validity divergence. Exploitation is further limited to networks/nodes still running Zebra < 4.4.0 and processing V5 transactions.

This is a consensus-divergence (chain-split) flaw in the Zebra Zcash node, scored CVSS 4.0 8.3 with VI:H/VA:H and VC:N - meaning the real impact is on integrity and availability of blockchain consensus, NOT confidentiality (which directly contradicts the 'Information Disclosure' input tag). No EPSS score, no CISA KEV listing, and no public POC were supplied, so there is no public exploit identified at time of analysis and no evidence of active exploitation. The CVSS 4.0 vector sets AT:P (attack requirement present), which is the key qualifier: an attacker can trivially craft and broadcast the malformed transaction (AV:N, AC:L, PR:N, UI:N), but the damaging chain-split only materializes if a Zebra-based miner selects it via getblocktemplate and mines it into a block that zcashd then rejects. Weighing the signals: the vector numbers are genuinely serious for a cryptocurrency network (forks, transaction-finality disruption, denial of service for Zebra-mined chains), but realized impact is gated on miner behavior rather than being a direct, self-contained compromise. A fix exists (4.4.0), reported by VulnCheck and confirmed with source-level analysis by the Zcash Foundation. Net assessment: a credible, real priority for Zcash node/miner operators to patch, but its narrow deployment footprint and dependency on a Zebra miner mining the crafted transaction temper the practical urgency relative to the raw 8.3 score.

An attacker crafts a malformed V5 transparent transaction containing a SIGHASH_SINGLE input whose index has no corresponding output (fewer outputs than inputs) and broadcasts it to the Zcash network. Zebra computes a ZIP-244 digest and accepts the transaction into its mempool instead of rejecting it, and a Zebra-based miner then templates it via getblocktemplate and mines it into a block that zcashd nodes reject, splitting the chain. No public exploit code is identified at time of analysis, though the Zcash Foundation advisory documents the exact divergent code paths.