This week’s newsletter features draft Bitcoin Improvement Proposals (BIPs) aimed at enhancing exfiltration-resistant signing, summarizes the identification of three vulnerabilities in older Bitcoin Core versions, outlines plans for mining the testnet5 genesis block, and discusses an ongoing investigation into issues with compact block reconstruction. Regular sections also announce new releases and candidates, highlighting significant updates to popular Bitcoin infrastructure software.
News
Several important updates have emerged in the Bitcoin community, including proposals for enhanced security measures and the identification of vulnerabilities in Bitcoin Core.
- Draft BIPs for anti-exfil signing:Connor Avila shared a proposal on the Bitcoin-Dev mailing list to standardize exfiltration-resistant signing for hardware devices. Due to nonce selection by the signing device, a compromised device could potentially leak its seed. The proposal suggests allowing the host to contribute to nonce generation, a method already adopted by various manufacturers but lacking standardization (see Newsletter #136). This discussion is not new; Stepan Snigirev proposed a similar standardization in 2020 (see Newsletter #87), while Pieter Wuille provided an overview of various techniques (see Newsletter #88). In 2024, Moonsettler introduced an anti-exfiltration protocol (see Newsletter #317), and the recently merged BIP461 facilitates the identification of potential key material leaks from signing devices (see Newsletter #424). Avila’s proposal includes two draft BIPs: the first addresses anti-exfil signing for Schnorr and ECDSA signatures, outlining a two-round protocol where the host introduces its own randomness for nonce generation and verifies its use by the signing device. The second draft incorporates three specific per-input fields into Partially Signed Bitcoin Transactions (PSBTs) to execute the anti-exfil protocol. This proposal also provides test vectors, a Python reference implementation, and a C module for libsecp256k1. Avila mentioned that Sjors Provoost has already released an experimental draft for single-signer taproot key-path signing and plans to synchronize his implementations with Avila’s specifications. Lastly, Avila has invited feedback and reviews from other developers, noting that his proposal has not yet undergone a cryptographic audit.
- Disclosure of three vulnerabilities in Bitcoin Core:Antoine Poinsot posted three security advisories on the Bitcoin-Dev mailing list concerning vulnerabilities in Bitcoin Core, one classified as high severity and two as medium severity, which were addressed in releases 29.2 and 30.0. According to the disclosure policy (see Newsletter #306), vulnerabilities of medium and high severity are reported two weeks after the last affected release expires, which for the 29.x series occurred on October 27. The vulnerabilities disclosed are:
- Block template creation halted by TRUC sibling eviction during reorg:This high-severity bug impacted mining nodes, preventing block template creation after specific reorganizations until a restart. TRUC sibling eviction continued to function when transactions from disconnected blocks returned to the mempool, potentially evicting a transaction while its descendants remained. This issue affected versions 28.0 through 29.1, was responsibly disclosed by Niklas Gögge in September 2025, and fixed by Greg Sanders in Bitcoin Core #33504 (see Newsletter #375).
- Inbound Tor connections wrongly receive -whitelist permissions:This medium-severity bug could remove DoS limits and privacy safeguards for all inbound Tor peers on nodes with a -whitelist range that included their Tor daemon's address, a configuration used by many reachable nodes. Martin Zumsande and Antoine Poinsot responsibly disclosed the issue in September 2025, with Zumsande addressing it in Bitcoin Core #33395.
- Block processing stall via witness malleation:A medium-severity bug allowed a peer that first announced a compact block to stall a node’s processing of that block for up to 10 minutes, with the possibility of indefinite stalling under certain conditions. Eugene Siegel responsibly disclosed this bug in April 2025, and it was resolved by Greg Sanders in Bitcoin Core #32646. Fixes for the first two vulnerabilities were also included in Bitcoin Core 28.3. All affected release series have now reached their end of life, so users are urged to update to a supported release.
- Mining plan for testnet5 genesis block:Anthony Towns shared a strategy on the Bitcoin-Dev mailing list to mine the genesis block of testnet5, which is proposed to replace testnet4 as outlined in BIP95 (see Newsletters #409 and #412). Mining is scheduled to commence once the mainnet undergoes its next difficulty adjustment at block 975,744, anticipated on November 13. The genesis block will commit to a mainnet block hash at or after that height, preventing any advance mining of testnet5 blocks. Given that testnet5’s minimum difficulty is approximately a million times higher than that of testnet4 (to avert a rapid influx of blocks post-launch before difficulty adjusts), the initial blocks will necessitate ASICs. Towns noted a public pool will be available for anyone to direct hashrate towards, ensuring that developers do not gain an early advantage. Once the genesis block is mined, its values will replace the placeholders in BIP95. Towns also pointed out that node software will require testnet5 parameters, BIP54 consensus cleanup logic, as well as DNS seeds, faucets, and explorers, before the network becomes available for general use.
- Ongoing discussion on compact block reconstruction statistics:0xB10C contributed to his Delving Bitcoin thread (see Newsletter #382) by explaining the poor performance of compact block relay observed from mid-July to mid-October. In September, he noted that his nodes could reconstruct only 1% to 5% of blocks without requesting transactions, resulting in approximately 100 kB per block being requested. Gregory Sanders discovered that nearly all missing transactions belonged to chains of rune mints, which reached mining pools moments before they were available to other nodes. Research on the Bitcoin Network Operations Collective forum (one, two) traced these chains back to an individual minting tokens as inexpensively as possible. Each transaction was about 100 vbytes, included a pay-to-anchor (P2A) input along with P2A and OP_RETURN outputs, and paid around 10 sat in fees. Peers sent the chains directly to nodes without announcing them, while other peers transmitted transactions that spent a P2A output from the middle of a chain, paying its entire value as fees. The choice between retaining the chain or its replacement depended partly on which arrived first. Consequently, mempools diverged from the sets mined, necessitating requests for the missing transactions.
Releases and release candidates
New releases and release candidates are available for notable Bitcoin infrastructure projects. Users are encouraged to upgrade to new releases or assist in testing release candidates.
- LDK v0.3-rc4:This is the fourth release candidate for the upcoming major version of this library designed for building Lightning Network-enabled wallets and applications, following the candidate detailed in Newsletter #426. Developers should examine the API and backward compatibility changes before testing, particularly noting the invalidation of previously issued BOLT11 invoices containing payment metadata.
- LDK v0.2.8 and v0.1.14:These are security releases for the 0.2 and 0.1 branches of this library, respectively. Both address vulnerabilities that could result in fund loss or denial-of-service attacks, along with cooperative-close issues mentioned below and a route-hint handling issue. Version 0.2.8 also resolves a fee overpayment vulnerability affecting static BOLT12 invoice payments and denial-of-service vulnerabilities relating to LSPS2 services.
- LND v0.21.8-beta and v0.20.8-beta:These are maintenance releases for this Lightning Network node implementation. Both contain fixes for AMP recovery, channel funding, restored-channel force-close, and confirmation-hint issues detailed below. They also enhance recovery of forwarded HTLCs amid link reconnections and include sweeper and channel state fixes.
Notable code and documentation changes
Recent significant updates have been made in various Bitcoin-related projects, including Bitcoin Core, Core Lightning, Eclair, LDK, LND, libsecp256k1, Hardware Wallet Interface (HWI), Rust Bitcoin, BTCPay Server, BDK, Bitcoin Improvement Proposals (BIPs), Lightning BOLTs, Lightning BLIPs, Bitcoin Inquisition, and BINANAs.
- Bitcoin Core #36142:This update rectifies command injection vulnerabilities through the Tor control endpoint. An attacker who controls or impersonates that endpoint could return a malicious onion service private key. Bitcoin Core now validates keys before saving or reusing them and checks returned onion service IDs. If a cached key fails validation, the onion service will be unavailable. Deleting the cached key and restarting will generate a new onion address.
- BIPs #2251:This adds BIP394, a draft specification for the rawtr() output script descriptor previously discussed in Newsletter #419. Unlike tr(), which derives a taproot output key from an internal key and an optional script tree, rawtr() takes the output key directly without applying a tweak. This enables a wallet to describe an output without understanding its internal structure; however, it does not prove the absence of other spending paths. Therefore, software importing from untrusted sources should alert users to the potential for hidden script paths. Bitcoin Core has supported this descriptor since version 24.0 (see Newsletter #213).
- BIPs #2347:This updates BIP95, the draft specification for testnet5 (see Newsletter #412), to simplify its proposed genesis block mining using pool infrastructure (see the news item above). The commitment to a recent mainnet block hash transfers from a public-key field in the coinbase output to readable text in the coinbase input script. The output uses OP_RETURN instead, marking it as unspendable and providing extra nonce space. The draft also allows miners to modify the block version, expanding the available mining search space. The example genesis block remains a placeholder.
- BIPs #2342:This clarifies BIP34’s encoding of the block height at the start of the coinbase input script for chains enforcing this rule immediately after genesis, such as signet and testnet4. Heights 1 to 16 must utilize single-byte opcodes OP_1 through OP_16, without a length prefix. The previous wording could be misinterpreted as necessitating a length-prefixed number. This clarification aligns with the existing behavior of Bitcoin Core and states that the coinbase input script must still contain at least two bytes, meaning the height opcode alone is insufficient (see Newsletter #406).
- Core Lightning #9610:This resolves a crash triggered by a peer due to an oversized HTLC amount. A channel peer could provoke this crash by proposing an HTLC with a fictitious amount of at least about 92 million BTC, setting the highest bit of the amount field. This could be interpreted as negative during a signed balance check, bypassing the necessary funds check. Subsequent commitment transaction construction led to an assertion failure, terminating the affected channel’s subprocess. Now, CLN rejects amounts exceeding the BOLT1 ceiling of 21 million BTC, expressed in millisatoshis, before executing balance arithmetic.
- LDK #5507 and #5532:These address cooperative-close bugs that could incorrectly combine signatures for different transactions. The first ensures that LDK signs the peer’s proposed transaction rather than reusing a cached signature from another closing transaction because the proposed fee matched. The second PR resets the pending signing state when a peer disconnects after an unavailable async signer, preventing an old peer signature from being inappropriately combined with a new local proposal after reconnection. In both scenarios, LDK could erroneously submit an invalid closing transaction for broadcast while treating the channel as closed, risking fund recovery.
- LND #11492:This resolves issues with on-chain recovery of received atomic multipath payments (AMP) post-force close. Previously, LND might fail to locate the preimage necessary to claim an outstanding HTLC, despite having recorded the payment as settled. LND incorrectly checked for an invoice-level preimage instead of the preimage stored for the individual AMP HTLC. It now correctly retrieves that HTLC’s preimage. The PR also prevents expiration checks intended for new payments from rejecting previously recorded spontaneous AMP and keysend HTLCs during recovery (see Newsletter #412 for earlier details on LND’s expiration checks).
- LND #11479:This fixes the error handling when publishing a transaction that opens multiple channels (see Newsletter #166). In the past, an error during broadcast caused LND to abandon pending channels and release their reserved wallet inputs, even if the transaction had already reached the network and could still confirm. Now, LND maintains the pending channels’ state once publication is attempted, preserving the necessary information to monitor the transaction and recover the channels. If the transaction fails to confirm, the retained pending channels will require manual cleanup.
- LND #11340 and #11374:These extend LND’s experimental BOLT12 offers support (see Newsletter #424) with offer creation and invoice request handling. The first PR introduces a handler that validates requests against stored offers and responds with signed invoices. The second PR connects the handler to incoming onion messages and adds the CreateOffer RPC and the lncli createoffer command. However, the handler currently does not retain the invoice state needed for payment settlements against the invoices it issues. This feature is still in development builds and requires further configuration.
- LND #11480:This resolves a crash when attempting to force close a channel restored from a static channel backup. Such backups do not contain a usable local commitment transaction; recovery necessitates the counterparty publishing its commitment. Previously, the force-close process checked for local data loss but did not account for the restored-channel status, which could lead to accessing a missing commitment transaction and causing panic. LND now appropriately rejects local force-close attempts and handles unavailable transactions safely, even when resuming an interrupted close after a restart.
- LND #11481:This resolves how LND caches the starting block height for scans verifying transaction confirmations. An attacker could propose a fake channel funding output with the txid of a revoked commitment transaction and a different output script. Since LND cached scan starting heights by txid alone, the fake funding request shared its cache entry with the revoked commitment request. This could advance the starting height past the revoked transaction’s confirmation block, delaying LND’s detection of the revoked commitment and its penalty transaction. LND now maintains separate starting heights for each txid and output script. This issue affects historical scans, including those used by Neutrino and Bitcoin Core without a transaction index.