ZEBRA is a Zcash node written entirely in Rust. Prior to zebrad version 4.3.1 and prior to zebra-script version 5.0.2, after a refactoring, Zebra failed to validate a consensus rule that restricted the possible values of sighash hash types for V5 transactions which were enabled in the NU5 network upgrade. Zebra nodes could thus accept and eventually mine a block that would be considered invalid by zcashd nodes, creating a consensus split between Zebra and zcashd nodes. In a similar vein, for V4 transactions, Zebra mistakenly used the "canonical" hash type when computing the sighash while zcashd (correctly per the spec) uses the raw value, which could also crate a consensus split. This issue has been patched in zebrad version 4.3.1 and zebra-script version 5.0.2.

Zebra (zebrad) before 6.2.1 contains an asymmetric resource consumption vulnerability that allows unauthenticated peers to stall block verification by pushing V6 mempool transactions with invalid Halo2 proofs. Attackers can flood the shared unprioritized Halo2 verification queue with zero-fee transactions carrying zero-filled Orchard and Ironwood proofs, causing nodes to fall behind the chain tip.

Zebra zebrad 4.5.0 and zebra-script 7.0.0 count P2SH redeem script signature operations in legacy mode rather than zcashd's accurate P2SH mode, overcounting CHECKMULTISIG preceded by OP1 through OP16 as 20 sigops and causing a consensus divergence. Remote attackers can broadcast P2SH spends using low-threshold multisig redeem scripts so that a block zcashd accepts exceeds Zebra's inflated MAXBLOCKSIGOPS count, causing Zebra nodes to reject it and stall off the chain.

The block sync download path in Zebra (zebrad) before 6.3.0 reads a block's height from its unvalidated coinbase scriptSig and drops blocks that appear too far behind the tip before consensus validation, without penalizing the supplying peer. Because V5 transaction IDs exclude the scriptSig, a malicious peer can repeatedly serve a canonical block whose coinbase claims height 1 while keeping the requested hash, delaying the node's discovery of the newest block.

Zebra zebrad 4.4.0 and zebra-script 6.0.0 fail to enforce a ZIP-244 consensus rule, accepting V5 transparent inputs signed with SIGHASHSINGLE that lack a corresponding output. Attackers can broadcast crafted V5 transactions with more inputs than outputs that Zebra accepts but zcashd rejects, causing a network consensus split.

Zebra (zebrad) 5.0.0 before 6.0.0-rc.0 does not apply its per-peer mempool admission cap to transactions received as direct P2P tx messages, because these are queued without the sending peer recorded as their source. A remote inbound peer can push many unique transactions to occupy a disproportionate share of mempool admission slots, crowding out honest peers' transaction relay.

Zebra (zebrad) 4.5.0 before 6.3.0 discards which peer supplied the block hashes in FindBlocks responses, then assigns 100 misbehavior points, the ban threshold, to whichever peer serves a requested block more than 50,000 heights above the tip. A remote peer can return real far-ahead hashes to a syncing node so that honest peers get banned, eroding its peer set and raising eclipse risk.