BitcoinCVEs & Vulnerabilities

Bitcoin Core before 24.1, when debug mode is not used, allows attackers to cause a denial of service (e.g., CPU consumption) because draining the inventory-to-send queue is inefficient, as exploited in the wild in May 2023.

In Bitcoin Core through 26.0 and Bitcoin Knots before 25.1.knots20231115, datacarrier size limits can be bypassed by obfuscating data as code (e.g., with OP_FALSE OP_IF), as exploited in the wild by Inscriptions in 2022 and 2023. NOTE: although this is a vulnerability from the perspective of the Bitcoin Knots project, some others consider it "not a bug."

Bitcoin Core 0.13.0 through 29.x has an integer overflow.

Bitcoin Core through 29.0 allows a denial of service via a crafted transaction.

Bitcoin Core through 29.0 allows Uncontrolled Resource Consumption (issue 2 of 2).

Bitcoin Core through 29.0 allows Uncontrolled Resource Consumption (issue 1 of 2).

Bitcoin Core through 27.2 allows transaction-relay jamming via an off-chain protocol attack, a related issue to CVE-2024-52913. For example, the outcome of an HTLC (Hashed Timelock Contract) can be changed because a flood of transaction traffic prevents propagation of certain Lightning channel transactions.

In Bitcoin Core before 25.1, an attacker can cause a node to not download the latest block, because there can be minutes of delay when an announcing peer stalls instead of complying with the peer-to-peer protocol specification.

In Bitcoin Core before 25.0, a peer can affect the download state of other peers by sending a mutated block.

Bitcoin Core before 0.20.0 allows remote attackers to cause a denial of service (infinite loop) via a malformed GETDATA message.

Bitcoin Core before 22.0 has a CAddrMan nIdCount integer overflow and resultant assertion failure (and daemon exit) via a flood of addr messages.

Bitcoin Core before 22.0 has a miniupnp infinite loop in which it allocates memory on the basis of random data received over the network, e.g., large M-SEARCH replies from a fake UPnP device.

Bitcoin Core before 0.15.0 allows a denial of service (OOM kill of a daemon process) via a flood of minimum difficulty headers.

Bitcoin Core before 0.20.0 allows remote attackers to cause a denial of service (memory consumption) via a crafted INV message.

In Bitcoin Core before 0.18.0, a node could be stalled for hours when processing the orphans of a crafted unconfirmed transaction.

In Bitcoin Core before 0.21.0, an attacker could prevent a node from seeing a specific unconfirmed transaction, because transaction re-requests are mishandled.

Bitcoin Core before 0.21.0 allows a network split that is resultant from an integer overflow (calculating the time offset for newly connecting peers) and an abs64 logic bug.

Bitcoin Core before 24.0.1 allows remote attackers to cause a denial of service (daemon crash) via a flood of low-difficulty header chains (aka a "Chain Width Expansion" attack) because a node does not first verify that a presented chain has enough work before committing to store it.

Bitcoin Core before 25.0 allows remote attackers to cause a denial of service (blocktxn message-handling assertion and node exit) by including transactions in a blocktxn message that are not committed to in a block's merkle root. FillBlock can be called twice for one PartiallyDownloadedBlock instance.

Memory management and protection issues in Bitcoin Core v22 allows attackers to modify the stored sending address within the app's memory, potentially allowing them to redirect Bitcoin transactions to wallets of their own choosing.

Bitcoin Core 0.12.0 through 0.21.1 does not properly implement the replacement policy specified in BIP125, which makes it easier for attackers to trigger a loss of funds, or a denial of service attack against downstream projects such as Lightning network nodes. An unconfirmed child transaction with nSequence = 0xff_ff_ff_ff, spending an unconfirmed parent with nSequence <= 0xff_ff_ff_fd, should be replaceable because there is inherited signaling by the child transaction. However, the actual PreChecks implementation does not enforce this. Instead, mempool rejects the replacement attempt of the unconfirmed child transaction.

Bitcoin Core before 0.19.0 might allow remote attackers to execute arbitrary code when another application unsafely passes the -platformpluginpath argument to the bitcoin-qt program, as demonstrated by an x-scheme-handler/bitcoin handler for a .desktop file or a web browser. NOTE: the discoverer states "I believe that this vulnerability cannot actually be exploited."

bitcoind in Bitcoin Core through 0.21.0 can create a new file in an arbitrary directory (e.g., outside the ~/.bitcoin directory) via a dumpwallet RPC call. NOTE: this reportedly does not violate the security model of Bitcoin Core, but can violate the security model of a fork that has implemented dumpwallet restrictions

Bitcoin Core 0.20.0 allows remote denial of service.

Bitcoin Core 0.16.x before 0.16.2 and Bitcoin Knots 0.16.x before 0.16.2 allow remote denial of service via a flood of multiple transaction inv messages with random hashes, aka INVDoS. NOTE: this can also affect other cryptocurrencies, e.g., if they were forked from Bitcoin Core after 2017-11-15.

Bitcoin Core before 0.14 allows an attacker to create an ostensibly valid SPV proof for a payment to a victim who uses an SPV wallet, even if that payment did not actually occur. Completing the attack would cost more than a million dollars, and is relevant mainly only in situations where an autonomous system relies solely on an SPV proof for transactions of a greater dollar amount.

bitcoind and Bitcoin-Qt prior to 0.17.1 allow injection of arbitrary data into the debug log via an RPC call.

bitcoind and Bitcoin-Qt prior to 0.15.1 have a stack-based buffer overflow if an attacker-controlled SOCKS proxy server is used. This results from an integer signedness error when the proxy server responds with an acknowledgement of an unexpected target domain name.

bitcoind and Bitcoin-Qt prior to 0.10.2 allow attackers to cause a denial of service (disabled functionality such as a client application crash) via an "Easy" attack.

In Bitcoin Core 0.18.0, bitcoin-qt stores wallet.dat data unencrypted in memory. Upon a crash, it may dump a core file. If a user were to mishandle a core file, an attacker can reconstruct the user's wallet.dat file, including their private keys, via a grep "6231 0500" command.

Bitcoin Core 0.12.0 through 0.17.1 and Bitcoin Knots 0.12.0 through 0.17.x before 0.17.1.knots20181229 have Incorrect Access Control. Local users can exploit this to steal currency by binding the RPC IPv4 localhost port, and forwarding requests to the IPv6 localhost port.

Bitcoin Core 0.14.x before 0.14.3, 0.15.x before 0.15.2, and 0.16.x before 0.16.3 and Bitcoin Knots 0.14.x through 0.16.x before 0.16.3 allow a remote denial of service (application crash) exploitable by miners via duplicate input. An attacker can make bitcoind or Bitcoin-Qt crash.

In Bitcoin Core before v0.13.0, a non-final alert is able to block the special "final alert" (which is supposed to override all other alerts) because operations occur in the wrong order. This behavior occurs in the remote network alert system (deprecated since Q1 2016). This affects other uses of the codebase, such as Bitcoin Knots before v0.13.0.knots20160814 and many altcoins.

Bitcoin Core before v0.13.0 allows denial of service (memory exhaustion) triggered by the remote network alert system (deprecated since Q1 2016) if an attacker can sign a message with a certain private key that had been known by unintended actors, because of an infinitely sized map. This affects other uses of the codebase, such as Bitcoin Knots before v0.13.0.knots20160814 and many altcoins.

The Bitcoin Proof-of-Work algorithm does not consider a certain attack methodology related to 80-byte block headers with a variety of initial 64-byte chunks followed by the same 16-byte chunk, multiple candidate root values ending with the same 4 bytes, and calculations involving sqrt numbers. This violates the security assumptions of (1) the choice of input, outside of the dedicated nonce area, fed into the Proof-of-Work function should not change its difficulty to evaluate and (2) every Proof-of-Work function execution should be independent. NOTE: a number of persons feel that this methodology is a benign mining optimization, not a vulnerability

The Bloom Filter implementation in bitcoind and Bitcoin-Qt 0.8.x before 0.8.4rc1 allows remote attackers to cause a denial of service (divide-by-zero error and daemon crash) via a crafted sequence of messages.

Unspecified vulnerability in bitcoind and Bitcoin-Qt 0.8.x allows remote attackers to cause a denial of service (memory consumption) via a large amount of tx message data.

The HTTPAuthorized function in bitcoinrpc.cpp in bitcoind 0.8.1 provides information about authentication failure upon detecting the first incorrect byte of a password, which makes it easier for remote attackers to determine passwords via a timing side-channel attack.

bitcoind and Bitcoin-Qt before 0.4.9rc2, 0.5.x before 0.5.8rc2, 0.6.x before 0.6.5rc2, and 0.7.x before 0.7.3rc2, and wxBitcoin, do not properly consider whether a block's size could require an excessive number of database locks, which allows remote attackers to cause a denial of service (split) and enable certain double-spending capabilities via a large block that triggers incorrect Berkeley DB locking.

bitcoind and Bitcoin-Qt 0.8.x before 0.8.1 do not enforce a certain block protocol rule, which allows remote attackers to bypass intended access restrictions and conduct double-spending attacks via a large block that triggers incorrect Berkeley DB locking in older product versions.

The CTransaction::FetchInputs method in bitcoind and Bitcoin-Qt before 0.8.0rc1 copies transactions from disk to memory without incrementally checking for spent prevouts, which allows remote attackers to cause a denial of service (disk I/O consumption) via a Bitcoin transaction with many inputs corresponding to many different parts of the stored block chain.

bitcoind and Bitcoin-Qt 0.8.0 and earlier allow remote attackers to cause a denial of service (electricity consumption) by mining a block to create a nonstandard Bitcoin transaction containing multiple OP_CHECKSIG script opcodes.

bitcoind and Bitcoin-Qt before 0.4.9rc1, 0.5.x before 0.5.8rc1, 0.6.0 before 0.6.0.11rc1, 0.6.1 through 0.6.5 before 0.6.5rc1, and 0.7.x before 0.7.3rc1 make it easier for remote attackers to obtain potentially sensitive information about returned change by leveraging certain predictability in the outputs of a Bitcoin transaction.

The penny-flooding protection mechanism in the CTxMemPool::accept method in bitcoind and Bitcoin-Qt before 0.4.9rc1, 0.5.x before 0.5.8rc1, 0.6.0 before 0.6.0.11rc1, 0.6.1 through 0.6.5 before 0.6.5rc1, and 0.7.x before 0.7.3rc1 allows remote attackers to determine associations between wallet addresses and IP addresses via a series of large Bitcoin transactions with insufficient fees.

The alert functionality in bitcoind and Bitcoin-Qt before 0.7.0 supports different character representations of the same signature data, but relies on a hash of this signature, which allows remote attackers to cause a denial of service (resource consumption) via a valid modified signature for a circulating alert.

Unspecified vulnerability in bitcoind and Bitcoin-Qt allows attackers to cause a denial of service via unknown vectors, a different vulnerability than CVE-2012-4682.

Unspecified vulnerability in bitcoind and Bitcoin-Qt allows attackers to cause a denial of service via unknown vectors, a different vulnerability than CVE-2012-4683.

Unspecified vulnerability in bitcoind and Bitcoin-Qt before 0.4.7rc3, 0.5.x before 0.5.6rc3, 0.6.0.x before 0.6.0.9rc1, and 0.6.x before 0.6.3rc1 allows remote attackers to cause a denial of service (process hang) via unknown behavior on a Bitcoin network.