LL
← All projectsLightning Labs

LND

Go implementation of a complete Lightning Network node.

BitcoinLightning NetworkNormal
Repository coverage

1687 commits in the local evidence base

Every captured commit receives deterministic security triage and a separate communication-quality score. Security candidates and broader second-pass signals receive full-patch Ollama analysis.

186security candidates236second-pass queue1372AI analyses
162commits · 30 days
278commits · 60 days
643commits · 180 days
1422commits · 365 days
Backfill bands
Aug 5 → Feb 6405 seen5 candidatesComplete
Feb 6 → Jun 6665 seen28 candidatesComplete
Jun 6 → Jul 689 seen4 candidatesComplete
Jul 6 → Aug 569 seen0 candidatesComplete
Commit communication

Does the history explain itself?

Message quality measures whether a commit identifies its scope, purpose, rationale, testing, and supporting references. It does not change the security-severity score.

68/100 average clarity
387Strong · 80–100
837Adequate · 60–79
432Thin · 40–59
31Opaque · 0–39
1security candidate with opaque commit messaging
Read the scoring rubric →
Developer activity

Who is changing the project?

Public Git author strings; identities are not independently verified.

DeveloperCommitsCandidatesAnalyzedHigh riskMessage avg.
Olaoluwa Osuntokun29659235474
ziggie49922422265
Elle Mouton26832268167
Yong60926172
ziggieXXX50617170
George Tsagkarelis37335167
LNBiG111173
Jared Tobin909171
yyforyongyu1211574070
bitromortac731154071
Erick Cestari26726070
Suheb746074
Analysis record

Published AI watches

Last scanned 43 minutes ago

Low 43 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11588 from ziggie1984/lnwallet/fix-reorg-wallet-sync-flake

This change fixes a wallet synchronization check in LND. Previously, the wallet only compared block heights to decide if it was fully caught up with the Bitcoin chain. During a chain reorganization (a temporary fork where one block is repl…

State synchronization check previously used height only, which is insufficient across reorgsFix adds block hash comparison to prevent false-positive sync statusCould cause LND to proceed while wallet state reflects an invalidated or not-yet-updated chain tip
8cdbf85cby Yong+224−1354 files
Vendor flagged security relevance
Moderate 51 AI analysisMessage 68 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #10708 from jtobin/alias-perms

This commit fixes a missing permission check on a new Lightning routing RPC called XFindBaseLocalChanAlias. Without the entry, any valid macaroon could call that RPC, bypassing the normal access-control list. The patch adds the required pe…

Missing macaroon authorization entry for an RPC endpointPotential privilege escalation: any valid macaroon could invoke an unprotected RPCRegression test added to enforce macPermissions coverage
98dc2b44by Olaoluwa Osuntokun+35−02 files
No security note in commit
Moderate 68 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11555 from yyforyongyu/invoice-settlement-public

This large LND patch changes when an ordinary invoice is considered 'paid'. Previously, releasing the preimage (the secret that lets the payer claim the funds) immediately marked the invoice as settled. Now the invoice enters a new interme…

New intermediate invoice/HTLC state prevents premature settlement reportingOn-chain HTLC outcomes persisted without opt-in to support restart recoveryForwarding package marker persisted before preimage release to avoid replay rejection on restart
895118acby Olaoluwa Osuntokun+4536−32048 files
Vendor flagged security relevance
Informational 15 AI analysisMessage 80 · Strong
LL Lightning LabsLND BitcoinLightning Network

docs: describe invoice collection finality

This commit only adds documentation to the release notes. It describes a previously merged code change (PR 11555) about how LND handles invoice finality, but the commit itself contains no code, bug fixes, or behavior changes. There is noth…

9649d9f6by yyforyongyu+87−01 file
No security note in commit
Moderate 64 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11551 from ziggie1984/public-pr149-sanitized

This change fixes a bug in LND's invoice handling for multi-part payments (MPP and AMP). When a payment is split into several pieces (HTLCs), each piece is accepted individually. Previously, an old accepted piece could become too close to …

Fixes missing revalidation of accepted HTLC CLTV expiries at MPP/AMP set completionPrevents settlement of shards whose on-chain claim window has become unsafe while heldAdds new fail resolution ResultHtlcSetExpiryTooSoon and cancels the entire affected set
4b7447bcby Yong+281−347 files
Vendor flagged security relevance
Low 35 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11396 from yyforyongyu/task-lease-policy

This commit changes how LND's wallet locks (leases) unspent coins through its RPC interface. It removes a recently added 'release after N confirmations' option and replaces it with an explicit 'no expiry' flag. Old clients that still send …

Breaking API change that reserves removed protobuf fields to prevent accidental reintroductionExplicit rejection of legacy `release_after_spend_confs` parameters from old clients instead of silent fallbackValidation prevents conflicting no-expiry + duration lease settings before wallet state is touched
03e28a96by Yong+663−4029 files
Vendor flagged security relevance
Low 40 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11175 from ViktorT-11/2026-08-gossip-version-agnostic-store

This commit refactors how the Lightning Network Daemon (LND) stores and queries channel graph data so that the same channel or node can exist under both old (v1) and new (v2) gossip protocols without the database returning stale, missing, …

Cross-version data consistency: prevents a v2 channel shell from shadowing a usable v1 channel, which could affect routing decisions and RPC output.Cache consistency: ensures the in-memory pathfinding cache and the SQL-backed RPC reads agree on which channel version and policies are active.Bug fix: zombie edge info now carries the correct channel ID instead of 0, which could have caused incorrect channel identification during resurrection.
8c251687by Yong+3752−28726 files
No security note in commit
Moderate 61 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11440 from Roasbeef/fix-longstanding-correctness-bugs

This is a large bug-fix patch for LND (a Bitcoin Lightning Network node) that tightens correctness rules across many subsystems. It prevents several ways a node could mis-handle channel data, accept bad network gossip, commit corrupted dat…

Database transaction commit now aborts on read errors (etcd) - prevents silent data corruption / inconsistent stateGossip zombie marking removed for invalid funding script - prevents denial-of-service / channel graph poisoningWire message length bounding added to watchtower protocol - prevents memory exhaustion / malformed message parsing
0123a548by Olaoluwa Osuntokun+1424−11938 files
Vendor flagged security relevance
Moderate 59 AI analysisMessage 80 · Strong
LL Lightning LabsLND BitcoinLightning Network

invoices: require complete payment collection

This commit changes how LND handles multi-part Lightning payments when one part fails after the preimage was already released. Previously, the invoice could be marked as fully paid even if one shard was later reclaimed by the payer on-chai…

Changes invoice settlement semantics for released-but-failed HTLCsAdds new persisted HTLC state `HtlcStateFailed`Prevents invoice from reaching `ContractSettled` when any released shard fails
699ea930by Olaoluwa Osuntokun+492−4910 files
No security note in commit
Moderate 60 AI analysisMessage 78 · Adequate
LL Lightning LabsLND BitcoinLightning Network

lnwallet: reject a zero CSV delay

This change tightens the rules for opening Lightning payment channels by refusing a setting called the CSV delay when it is set to zero. The CSV delay is the waiting period a user must endure before reclaiming their own funds from a channe…

Rejects a zero CSV delay that would eliminate the breach-remedy windowCentralizes the check in VerifyConstraints so all channel types and directions are coveredAdds explicit error type and unit test for the new constraint
266eba7bby Olaoluwa Osuntokun+81−03 files
Vendor flagged security relevance
Moderate 58 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

lnwallet+chanstate: keep persisted HTLC data readable

This commit fixes two bugs in how Lightning Network channel data is saved to disk. First, it stops extra 'aux' signatures meant only for advanced taproot channels from being saved on ordinary channels, where they don't belong. Second, it p…

Persisted state could become unreadable due to writer/reader size mismatchAuxiliary signatures for taproot overlay channels were persisted on plain channel HTLCsFix adds explicit serialization bound to prevent unreadable channel records
43c74723by Olaoluwa Osuntokun+101−25 files
No security note in commit
Low 39 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11374 from bitromortac/20261016-c7e5-dispatch

This commit adds the first end-to-end BOLT 12 'Offers' flow to LND: a new dev-only RPC and CLI command to create offers, plus a dispatcher that listens for incoming onion-message invoice requests and replies with signed BOLT 12 invoices. I…

New network-facing dispatcher answers invoice requests from any peer; rate and concurrency limits are appliedBlinded-path PathID check prevents reply linking to self-created paths, but TODO notes the check should move once offers can have offer_pathsCreateOffer RPC is restricted to dev builds and requires offchain:write macaroon permission
a9d4ca3bby bitromortac+3113−7838 files
No security note in commit
Moderate 60 AI analysisMessage 78 · Adequate
LL Lightning LabsLND BitcoinLightning Network

invoices: reject settled AMP set shards

This change fixes a bug in LND's invoice handling for AMP (Atomic Multi-Path) payments. Previously, if someone tried to pay a fresh shard for an AMP set that was already fully paid, the invoice updater could return an internal error that w…

Denial-of-service vector: an internal updater error could force-fail a shared channel link, disrupting unrelated sibling HTLCsAMP invoice state machine edge case: reusable invoice stays open for other sets but must reject new shards for already-settled setsClean failure path introduced: ResultInvoiceNotOpen instead of error-driven link failure
99e3da37by yyforyongyu+49−02 files
No security note in commit
Moderate 59 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11282 from lightningnetwork/elle-base-branch-payment-service

This is a large feature merge for LND that adds a new optional RPC subsystem called switchrpc. It lets an external program hand a pre-built onion packet to LND for direct dispatch over a Lightning channel, instead of LND doing its own path…

New optional RPC surface (switchrpc) that accepts externally crafted onion packets and directly dispatches HTLCsIdempotency/deduplication added via InitAttempt durable write-ahead log and ErrPaymentIDAlreadyExistsStartup orphan cleanup fails half-open or pre-commit attempts to avoid hangs
11f79be0by Olaoluwa Osuntokun+11385−22458 files
Vendor flagged security relevance
Low 44 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11472 from yyforyongyu/task-sweep-rejections

This commit fixes a bug in LND's transaction sweeping logic. Previously, if the Bitcoin backend rejected a sweep transaction for a temporary or unknown reason (for example, the mempool was full, a network error occurred, or the rejection r…

Funds-availability / liveness fix: temporary or unknown mempool rejections could leave sweep inputs unspent.Error-handling reclassification: previously fatal errors now retryable.Retry metadata preservation: next fee rate and fee cap kept across attempts.
c5863e55by Yong+290−9720 files
No security note in commit
Low 32 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11292 from ViktorT-11/2026-10-harden-remote-signing-no-op

This change is a hardening fix for LND's remote-signing feature. Previously, if a node was configured to act as an outbound remote signer but was compiled without the required 'signrpc' and 'walletrpc' sub-servers, it would silently fall b…

silent-degradation-to-no-op preventedconfiguration-misconfiguration now surfaces as fatal errorbuild-tag-dependency explicitly validated at runtime
0c538e1fby Suheb+241−125 files
No security note in commit
Informational 15 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

invoices: fix test whitespace

This commit only fixes blank-line spacing in a single test file to satisfy lint checks. It adds one blank line after a function signature and removes one trailing blank line. There are no code behavior changes, no bug fixes, and no securit…

a67d07f4by yyforyongyu+1−11 file
No security note in commit
Low 31 AI analysisMessage 92 · Strong
LL Lightning LabsLND BitcoinLightning Network

build: pin the new lint job actions by commit

This change updates two GitHub Actions used in the project's automated lint workflow so they are locked to specific commit hashes instead of floating version tags. A floating tag could be silently repointed to a different, possibly malicio…

CI/CD supply-chain hardeningthird-party action pinning by commit hashprevention of tag-movement attacks
2a9f137dby Olaoluwa Osuntokun+4−21 file
Vendor flagged security relevance
Informational 15 AI analysisMessage 92 · Strong
LL Lightning LabsLND BitcoinLightning Network

build: fix the custom linter cache wiring for CI

This commit fixes the GitHub Actions CI setup for running a custom code linter. It makes sure Docker can read and write build caches correctly and that the built linter image is available for later steps. There is no user-facing security f…

6f665663by Olaoluwa Osuntokun+30−82 files
No security note in commit
Moderate 66 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11482 from Roasbeef/pub/unsigned-acked-wrap

This patch fixes a database bug in the Lightning Network Daemon (LND) where very long lists of unsigned, acknowledged channel updates could be silently truncated when saved to disk. Because the storage format used a 16-bit counter, lists l…

Integer truncation / wrap-around in serialization format (uint16 count)Silent data loss on restart due to undersized length fieldDefensive deserialization with count-modulo validation
7431093fby Olaoluwa Osuntokun+218−127 files
Vendor flagged security relevance
Repository ledger

Explore captured commits

Expand any commit for its author, full message, clarity score, changed files, triage signals, analysis, and source link.

AI review queuedMerge pull request #11474 from ziggie1984/itest-fix-outbound-rs-macaroon-flakeby Yong · 03b2b8d4 · Nov 3, 2026 · 1 fileMessage 83 · StrongInformational 14Details
Commit message · Yong

Merge pull request #11474 from ziggie1984/itest-fix-outbound-rs-macaroon-flake

itest: deflake outbound remote signer macaroon test

83/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Mentions testing or verification✓ Links an issue, advisory, or supporting reference
Why it was queued
signing or wallet pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Informational 14/100

This commit only adjusts timing settings in an integration test to reduce flaky failures on slow CI systems. It is not a security fix and does not change production code.

AI review queueditest: deflake outbound remote signer macaroon testby ziggie · 06a97396 · Nov 3, 2026 · 1 fileMessage 83 · StrongInformational 15Details
Commit message · ziggie

itest: deflake outbound remote signer macaroon test

testOutboundRSMacaroonEnforcement waits a fixed 15 seconds for the
watch-only node to write its invoice macaroon. On slow CI runners with a
remote database backend, opening the database can take longer than that,
and the test fails with a missing macaroon file.

Wait for the macaroon with the default timeout, which already accounts
for remote database backends.

The watch-only node starts waiting for the signer after writing its
macaroons, and the test still has to open a stream and start the signer
node before the signer connects. Raise the watch-only node's signer
connection timeout to the default timeout as well, so a slow runner
doesn't move the flake to that step.

83/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Mentions testing or verification
Why it was queued
signing or wallet pathsecond-pass: security-sensitive path
AI analysis · Informational 15/100

This commit only changes an integration test. It replaces hard-coded 15-second waits with longer default timeouts so the test doesn't randomly fail on slow CI machines. There is no change to production code or user-facing behavior, and no security issue is being fixed or introduced.

Security candidateci: bump gateway to v0.7.2 and skip comments from non-membersby litbot-9000 · e46a7503 · Nov 2, 2026 · 1 fileMessage 85 · StrongLow 25Details
Commit message · litbot-9000

ci: bump gateway to v0.7.2 and skip comments from non-members

Bump the gateway-action and runtime pins to v0.7.2 together: the v0.7.2
runtime needs the job token that gateway-action v0.7.2 passes. The release
limits /gateway explain to maintainers and the PR author, authorizes every
command before any credential is loaded, and adds repeat guards for
explain (per finding) and check (per HEAD). Refused commands get no reply;
the reason is in the workflow log.

Also skip comments from anyone who is neither a repo/org member nor the PR
author, so they never boot a runner. The runtime still checks every
command's permission itself.

85/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context
Why it was queued
credential or privilege statedocumentation-only discount
AI analysis · Low 25/100

This commit updates a GitHub Actions workflow that runs an AI code-review assistant called 'gateway' on pull request comments. It adds a pre-filter so only repository members or the pull request author can trigger the workflow, and it upgrades the gateway software to a new version that checks permissions before loading any credentials. The change is defensive: it reduces the chance that an outsider could waste compute resources or interact with the review bot, and it makes credential handling stricter. There is no direct evidence in the commit of a past security breach.

Security candidatebuild: pin btcwallet rejection mappingsby yyforyongyu · 8a69c380 · Nov 2, 2026 · 2 filesMessage 88 · StrongLow 26Details
Commit message · yyforyongyu

build: pin btcwallet rejection mappings

Use btcwallet v0.18.1 for the rejection mappings from #1464. Its
maintenance base avoids #1452's Neutrino upgrade and keeps the
existing migration library and RPC generator versions.

https://github.com/btcsuite/btcwallet/pull/1464
https://github.com/btcsuite/btcwallet/pull/1452

88/100 · StrongMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Links an issue, advisory, or supporting reference
Why it was queued
access control
AI analysis · Low 26/100

This commit updates LND's dependency on btcwallet from a pre-release development snapshot to the stable v0.18.1 release. The commit message says the goal is to adopt new 'rejection mappings' from a btcwallet pull request while avoiding an unrelated Neutrino upgrade. The change itself only modifies Go module files (go.mod and go.sum); it does not contain any actual code changes. Whether this is a security fix depends entirely on what the referenced btcwallet pull request #1464 contains, which is not shown in the supplied materials.

Security candidateinvoices: reject stale shards at set completionby ziggie · 3ba74925 · Nov 1, 2026 · 4 filesMessage 68 · AdequateModerate 62Details
Commit message · ziggie

invoices: reject stale shards at set completion

Revalidate every accepted MPP or AMP shard when a completing shard
arrives. If an existing shard no longer satisfies either CLTV delta,
cancel the accepted set, keep the invoice open, and reject the
completing shard without recording it.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
defensive validationboot or update path
AI analysis · Moderate 62/100

This change fixes a bug in how Lightning Network invoices handle multi-part payments (MPP/AMP). Previously, a payment shard accepted early could sit until the rest of the payment arrived. By then, its timeout might be too close to the current blockchain height to safely settle, yet the system would still complete the payment. Now, when the final shard arrives, every previously accepted shard is re-checked for a safe timeout. If any shard is too close to expiring, the whole payment set is cancelled, the invoice stays open, and the final shard is rejected. This prevents a situation where funds could be released for a payment that can no longer be claimed on-chain in time.

Security candidatescripts: make OTS output deterministic for identical inputsby Boris Nagaev · 147e2b03 · Oct 31, 2026 · 1 fileMessage 73 · AdequateInformational 18Details
Commit message · Boris Nagaev

scripts: make OTS output deterministic for identical inputs

Keep the original CLI, dependencies, output layout and uploader behavior.
Fix the OTS subprocess hash seed, locale and timezone so identical responses
are traversed consistently. Sort proofs and replacement commands, omit local
backend details and generation times, and normalize output permissions.

Keep relative pending ages and transient diagnostics on the terminal rather
than in comparable files. Preserve the pending-age exit status.

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
credential or privilege state
AI analysis · Informational 18/100

This commit changes a release-maintenance helper script so that running it twice with the same inputs produces byte-for-byte identical output files. It removes timestamps, local backend details, and non-deterministic ordering from the generated artifacts, and fixes environment settings that could make the OpenTimestamps client behave differently on different machines. These are build/reproducibility hardening changes, not a fix for an exploitable vulnerability in the Lightning node itself.

Security candidateinvoices: persist ordinary HTLC outcomesby yyforyongyu · b0a3e8ae · Oct 31, 2026 · 5 filesMessage 68 · AdequateLow 39Details
Commit message · yyforyongyu

invoices: persist ordinary HTLC outcomes

Discover pending ordinary children across parent states and apply final
outcomes inside the existing invoice transaction. Recompute successful
value and assign the first settlement index once, preserving duplicates.

Reuse the KV scan and SQL query updater. Keep AMP metadata writes on
their existing add-time path and leave ordinary producers unchanged.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
boot or update path
AI analysis · Low 39/100

This commit changes how the Lightning Network Daemon (LND) records the final result of ordinary payment attempts (HTLCs) for invoices. It adds a way to find invoices whose child payments are still awaiting a final outcome, and applies those outcomes inside the existing invoice update process. The change is a correctness/robustness improvement rather than an obvious security fix, but it touches settlement accounting and state transitions, which are security-sensitive in a payment system.

Security candidatemulti: defer ordinary invoice settlementby yyforyongyu · 78c092be · Oct 31, 2026 · 15 filesMessage 68 · AdequateModerate 61Details
Commit message · yyforyongyu

multi: defer ordinary invoice settlement

Release ordinary, MPP and hold preimages into PENDING_SETTLE, then
assign final credit and the settlement index through recorded HTLC
outcomes. Recover pending children beneath terminal parents without
allocating another index; let recorded failures override channel
inference and retry channel-only completion while running.

Switch ordinary replay, preimage lookup, expiry, notifications and RPC
projection together. Preserve the parent’s AMP settlement mechanisms and
controls; AMP finality remains deferred. Reuse existing integration
invoice waits rather than adding another harness API.

Cover partial credit, all-failed cancellation, recovery races, transient
writes and pending RPC timestamps using the existing backend fixtures.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
boot or update path
AI analysis · Moderate 61/100

This commit changes how LND marks invoices as 'paid.' Previously, when a payment arrived, LND immediately marked the invoice as fully settled. Now, for ordinary, multi-part, and hold invoices, LND first releases the payment preimage to the peer but keeps the invoice in a new 'pending settle' state. It only flips the invoice to 'settled' once the underlying HTLC outcomes are recorded as final. AMP invoices keep their old behavior for now. The change is designed to prevent double-counting, incorrect settlement indexes, and premature credit allocation if a payment later fails or is replayed.

Security candidatedocs: add v0.20.7 and v0.21.6 release notesby ziggie · 050603ed · Oct 30, 2026 · 3 filesMessage 88 · StrongHigh 79Details
Commit message · ziggie

docs: add v0.20.7 and v0.21.6 release notes

Port the final v0.20.7 and v0.21.6 release notes to master, including
the CVE-2026-98432 notice urging Linux operators to disable MPTCP and
install a patched kernel. Remove the MPTCP fragment, which both patch
releases now announce.

88/100 · StrongMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Links an issue, advisory, or supporting reference
Why it was queued
explicit security languagedocumentation-only discount
AI analysis · High 79/100

This commit adds release notes for LND versions 0.20.7 and 0.21.6, warning Linux users about a serious Linux kernel security flaw (CVE-2026-98432). The notes explain that recent LND versions accidentally exposed a feature called MPTCP on network ports, which attackers could use to remotely exploit the kernel. The release notes tell operators to upgrade LND, disable MPTCP on their server, and install a patched Linux kernel. The actual code change that disables MPTCP was done in a separate pull request; this commit only documents it and removes a draft release-note fragment.

AI review queuedMerge pull request #11151 from Roasbeef/reject-absurd-feeratesby Yong · 5176c13c · Oct 30, 2026 · 10 filesMessage 73 · AdequateModerate 64Details
Commit message · Yong

Merge pull request #11151 from Roasbeef/reject-absurd-feerates

funding+lnwallet: reject absurdly high channel feerates

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
signing or wallet pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Moderate 64/100

This commit adds two safety checks to LND's channel-opening process. When someone tries to open a Lightning channel with LND, LND now rejects the request if the proposed on-chain fee rate is absurdly high (more than 10 times LND's own fee estimate, with a minimum floor of 100 sat/vbyte), or if the initial balances are so low that neither party could actually use the channel. These checks prevent attackers or misconfigured peers from tying up LND's resources on channels that would be useless or economically harmful.

Security candidatechancloser: preserve cooperative close variantsby Olaoluwa Osuntokun · 5b8e6a15 · Oct 29, 2026 · 5 filesMessage 73 · AdequateModerate 59Details
Commit message · Olaoluwa Osuntokun

chancloser: preserve cooperative close variants

In this commit, we carry the selected closing-signature variant through the cooperative close state machine instead of reducing it to a single no-closee flag.

When a peer selects the closee-only transaction, we now omit the remote output while building and completing the local proposal, then reply with the matching signature field. This keeps both peers on the same transaction for legacy and taproot channels.

73/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Names security-relevant behavior explicitly
Why it was queued
signing boundarysigning or wallet path
AI analysis · Moderate 59/100

This commit fixes a bug in how the Lightning Network Daemon (LND) handles cooperative channel closes. Previously, when one peer chose a special close variant that omits the 'closee' output, the other peer would not correctly preserve that choice. This could cause both peers to build slightly different closing transactions and fail to agree on the final close. The patch makes the chosen close variant flow through the entire close process so both sides produce and sign the exact same transaction, for both older and newer taproot-style channels.

AI review queuedsweep: retry inconclusive spend checksby Olaoluwa Osuntokun · d32a9497 · Oct 29, 2026 · 5 filesMessage 68 · AdequateModerate 57Details
Commit message · Olaoluwa Osuntokun

sweep: retry inconclusive spend checks

In this commit, we distinguish a failed spend lookup from a completed
lookup that found no spender.

Notifier registration and closed-subscription errors no longer crash the
node or get treated as "not spent" by accident. When the missing input
path can't tell whether an input was spent, it now returns a retryable
result instead of a fatal one, and the sweeper keeps the input's existing
starting fee preference rather than replacing it with the result's zero
value.

During regular record processing, an inconclusive check is logged and
the record proceeds as if no input was spent. Skipping it instead would
stop fee bumps for a deadline sweep for as long as the check keeps
failing. If an input was in fact spent, the mempool rejects the attempt
for missing inputs and the missing input handling takes over.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
second-pass: broader security terminology
AI analysis · Moderate 57/100

This commit fixes a bug in LND's transaction sweeping logic. Previously, when the node couldn't reliably check whether an input had already been spent (for example, because a notifier registration failed or a subscription closed unexpectedly), the code could either crash the node or wrongly treat the input as unspent. The patch makes these failures retryable instead of fatal, preserves the user's chosen fee rate instead of resetting it to zero, and continues fee-bumping deadline sweeps even when the spend check is inconclusive. This prevents stuck funds or incorrect fee handling during network or backend hiccups.

AI review queuedMerge pull request #11365 from saubyk/7316-listchaintxns-reverseby ziggieXXX · 730b9e55 · Oct 29, 2026 · 11 filesMessage 73 · AdequateInformational 20Details
Commit message · ziggieXXX

Merge pull request #11365 from saubyk/7316-listchaintxns-reverse

multi: add reverse ordering to listchaintxns

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
signing or wallet pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Informational 20/100

This commit adds a new optional 'reverse' flag to the lncli listchaintxns and GetTransactions RPC that lets users list wallet transactions oldest-first instead of newest-first. It also fixes a minor ordering stability issue so transactions in the same block now appear in a predictable order. There is no security vulnerability here; it is a user-facing feature and small bug-fix for transaction listing.

Security candidateMerge pull request #11431 from Roasbeef/actor-try-sendby Olaoluwa Osuntokun · c03d3659 · Oct 29, 2026 · 14 filesMessage 78 · AdequateLow 42Details
Commit message · Olaoluwa Osuntokun

Merge pull request #11431 from Roasbeef/actor-try-send

actor: add non-blocking TryTell/TryAsk, in-turn overflow sends, AskThen, and a shutdown join

78/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference✓ Names security-relevant behavior explicitly
Why it was queued
memory safetymerge-commit duplicate discount
AI analysis · Low 42/100

This commit adds new non-blocking message-sending tools and a bounded overflow queue to LND's internal actor framework. The goal is to prevent actor-to-actor message cycles from deadlocking when mailboxes are full, and to give callers a way to send messages without blocking. It also adds shutdown helpers so a caller can wait until an actor's current work finishes before cleaning up resources. The changes are defensive: they make overload and termination explicit (returning errors instead of silently blocking or dropping messages) and add hard caps so memory cannot grow without bound.

Security candidatedocs: add release notes for the actor non-blocking sendsby Olaoluwa Osuntokun · 8576d9f9 · Oct 28, 2026 · 1 fileMessage 90 · StrongInformational 15Details
Commit message · Olaoluwa Osuntokun

docs: add release notes for the actor non-blocking sends

In this commit, we add a release note for the non-blocking send, overflow,
AskThen, and shutdown join additions to the actor module.

90/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides detailed explanatory context✓ Names security-relevant behavior explicitly
Why it was queued
memory safetydocumentation-only discount
AI analysis · Informational 15/100

This commit only adds documentation (release notes) describing previously merged improvements to an internal actor messaging module. It contains no code changes and no security fixes or vulnerabilities.

Security candidateactor: add AskThen to receive an Ask reply as a later messageby Olaoluwa Osuntokun · a4eb9c7d · Oct 28, 2026 · 2 filesMessage 78 · AdequateInformational 24Details
Commit message · Olaoluwa Osuntokun

actor: add AskThen to receive an Ask reply as a later message

In this commit, we add AskThen, the pipe-to-self form of Ask. It sends
the request, returns immediately, and later delivers wrap(result) to the
caller's own ref as an ordinary message. A behavior handles the reply in
a later turn instead of calling Await inside Receive, which deadlocks
when the callee is, transitively, waiting on the caller.

AskThen never blocks its calling goroutine. Inside a receive turn the
request takes the in-turn send path, so a full target mailbox spills to
the overflow queue. From any other goroutine it is a TryAsk, so a full
mailbox is delivered to self as an ErrMailboxFull result.

Completion is best effort and detached, but never silently lost. The
result is retained and its delivery retried with TryTell while self is
alive and busy. It is dropped only when self has stopped, and then the
drop is logged, the message goes to the DLO, and an optional
WithDropHandler is called, so the caller can reconcile. A timeout
abandons the wait rather than the request, and the doc comment says that
a caller that needs a durable outcome must reconcile instead of relying
on the reply.

The design follows AskThen in wavelength's baselib/actor.

78/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Names security-relevant behavior explicitly
Why it was queued
memory safety
AI analysis · Informational 24/100

This commit adds a new helper called AskThen to LND's internal actor framework. It lets one actor send a request to another and receive the reply later as a normal message, instead of blocking and waiting. The main goal is to prevent deadlocks when two actors are waiting on each other. The change is defensive and includes tests, but it introduces new concurrency behavior that could affect reliability if misused.

Security candidateactor: never park sends made from inside a receive turnby Olaoluwa Osuntokun · db6106a3 · Oct 28, 2026 · 4 filesMessage 90 · StrongModerate 53Details
Commit message · Olaoluwa Osuntokun

actor: never park sends made from inside a receive turn

In this commit, we make a send issued from an actor's own receive turn
non-blocking. Parking an actor on a full mailbox stalls its whole
mailbox, and a cycle of actors that do so deadlocks.

The actor loop now stamps the context handed to the behavior with a
turn marker, which is cleared when the turn returns. A Tell or Ask made
with such a context goes through a new in-turn path on the default
ChannelMailbox: it uses the channel when there is room and nothing is
queued ahead of it, and otherwise spills into a bounded overflow queue
that the receive loop refills the channel from, oldest first.

The queue is capped at ActorConfig.MailboxOverflowLimit, defaulting to
DefaultMailboxOverflowLimit. A send at the cap fails with
ErrMailboxOverflow, which an Ask surfaces on its Future. Tell can't
return an error, so it logs the drop. A context that outlives its turn,
in a goroutine the behavior spawned, parks like any other producer.
WithoutTurn clears the marker explicitly.

Delivery order is acceptance order. TrySend reports full while the
overflow is non-empty, and an ordinary blocking Send from outside the
actor first waits for the overflow to drain instead of writing to the
channel ahead of it. An owner can therefore enqueue a shutdown or
barrier message and know it runs after everything already accepted.
Drain yields the overflow after the channel.

The Mailbox interface is unchanged: the in-turn path is an optional
interface that ChannelMailbox implements.

This is ported from the in-turn send design in wavelength's
baselib/actor, with the Send ordering added.

90/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Explains rationale or failure mode✓ Names security-relevant behavior explicitly
Why it was queued
memory safety
AI analysis · Moderate 53/100

This commit fixes a design flaw in LND's internal actor framework where an actor could freeze itself by sending a message to another actor whose mailbox was full. If several actors did this in a circle, the whole group could deadlock. The fix marks messages sent during an actor's own processing turn as 'in-turn' and routes them through a special non-blocking overflow queue, so the actor never stops moving. It also preserves message order and adds a hard cap so runaway message floods fail loudly instead of consuming unlimited memory.

AI review queuedMerge pull request #11357 from Roasbeef/zpay32-bolt11-vectorsby Olaoluwa Osuntokun · 5617ee7e · Oct 28, 2026 · 7 filesMessage 73 · AdequateLow 48Details
Commit message · Olaoluwa Osuntokun

Merge pull request #11357 from Roasbeef/zpay32-bolt11-vectors

zpay32: enforce BOLT 11 fixed lengths, single s, low-S with n

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
parser or protocol pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Low 48/100

This change tightens how LND reads BOLT 11 Lightning invoices. It now rejects invoices that violate the spec: fixed-length fields with wrong lengths, duplicate payment-address fields, and high-S signatures when the invoice explicitly names the payee. Previously these malformed invoices were accepted or silently skipped. The patch is defensive and brings LND into line with the BOLT 11 specification and its official test vectors.

AI review queuedzpay32: document first-wins for repeated n and h fieldsby Olaoluwa Osuntokun · b4c0fc91 · Oct 27, 2026 · 1 fileMessage 85 · StrongInformational 15Details
Commit message · Olaoluwa Osuntokun

zpay32: document first-wins for repeated n and h fields

In this commit, we update the comments on repeated `n` and `h` fields.
A field with the wrong length is now an error rather than unsupported,
so "already seen a supported one" no longer describes why we skip it.
Keeping the first one is deliberate: BOLT 11 only requires exactly one
`p` and one `s` field.

85/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Explains rationale or failure mode
Why it was queued
parser or protocol pathsecond-pass: security-sensitive path
AI analysis · Informational 15/100

This commit only changes code comments in the zpay32/decode.go file. It updates the explanation for why repeated 'n' and 'h' fields in BOLT 11 invoices are ignored after the first valid one. No actual code behavior was changed, so there is no security impact.

Security candidatezpay32: resync BOLT 11 vectors with bolts#1368by Olaoluwa Osuntokun · 8f72c8f0 · Oct 27, 2026 · 2 filesMessage 81 · StrongInformational 20Details
Commit message · Olaoluwa Osuntokun

zpay32: resync BOLT 11 vectors with bolts#1368

In this commit, we pick up the corrected high-S vector from
lightning/bolts#1368, which drops the duplicate `s` field it used to
carry. The vector now reaches the signature check, so we expect
`ErrHighSSignature` for it rather than `ErrDuplicatePaymentAddr`.

81/100 · StrongMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Links an issue, advisory, or supporting reference✓ Names security-relevant behavior explicitly
Why it was queued
signing boundary
AI analysis · Informational 20/100

This commit updates LND's test data and test expectations to match a corrected official BOLT 11 specification test vector. The invoice string used to exercise 'high-S' signature rejection was changed so it no longer contains a duplicate 's' field. The test now expects the code to reject it specifically for the high-S signature, not for a duplicate field. It is a test-only alignment with the spec, not a functional code change.

Security candidatezpay32: assert why each fixed BOLT 11 vector is rejectedby Olaoluwa Osuntokun · 79c123b0 · Oct 27, 2026 · 3 filesMessage 100 · StrongInformational 15Details
Commit message · Olaoluwa Osuntokun

zpay32: assert why each fixed BOLT 11 vector is rejected

In this commit, we make the tests show that each of the new decoder
rules is what rejects its invalid invoice, rather than any error.

The vector test only checked that an invalid vector failed, so a vector
rejected for the wrong reason still passed. That hid a real gap: the
spec's high-S vector also repeats its `s` field, so it's rejected as a
duplicate payment address before the signature is ever checked, and no
test exercised `ErrHighSSignature` at all.

We now map each vector this PR fixes to the sentinel error it must
produce, and add `TestDecodeHighSWithDestination`, which decodes the
same invoice (with a single `s` field) once with a high-S and once with
a low-S signature. A duplicate `s` case joins `TestParseTaggedFields`.
Dropping or breaking any of the three new checks now fails a test aimed
at it.

100/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Explains rationale or failure mode✓ Mentions testing or verification✓ Names security-relevant behavior explicitly
Why it was queued
signing boundary
AI analysis · Informational 15/100

This commit only improves and expands test code for the BOLT 11 invoice decoder. It makes existing tests verify the exact reason each invalid test invoice is rejected, and adds a new test specifically for high-S signature rejection. No production code is changed, so it cannot introduce or fix a runtime security vulnerability on its own.

AI review queuedMerge pull request #11354 from Roasbeef/taproot-vectors-bolts-1352by Olaoluwa Osuntokun · 4a4c9a5d · Oct 27, 2026 · 2 filesMessage 83 · StrongInformational 15Details
Commit message · Olaoluwa Osuntokun

Merge pull request #11354 from Roasbeef/taproot-vectors-bolts-1352

lnwallet: sync taproot test vectors with bolts#1352

83/100 · StrongMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Mentions testing or verification✓ Links an issue, advisory, or supporting reference
Why it was queued
fuzzing or regression evidencesigning or wallet pathmerge-commit duplicate discountsecond-pass: security-sensitive path
AI analysis · Informational 15/100

This commit only updates test data and test code for Taproot channel scripts in LND. It adds new test vectors that match the latest BOLT specification draft (bolts#1352), including separate keys for the remote party's commitment and a corrected distinction between the combined MuSig2 funding key and the final BIP-86 tweaked funding key. No production wallet, consensus, or networking code is changed, so this cannot directly affect live funds or node security.

AI review queuedMerge pull request #11287 from ziggie1984/fix-submodule-lintby Yong · 1a359d4a · Oct 23, 2026 · 57 filesMessage 73 · AdequateInformational 19Details
Commit message · Yong

Merge pull request #11287 from ziggie1984/fix-submodule-lint

multi: fix submodule lint issues and lint submodules in CI

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body✓ Links an issue, advisory, or supporting reference
Why it was queued
merge-commit duplicate discountsecond-pass: unusually broad change
AI analysis · Informational 19/100

This is a large housekeeping commit that fixes code-style and lint warnings across many submodules and adds a new CI step to lint those submodules. The only functional code change is a small rewrite of the retry-delay calculation to avoid integer overflow when retry settings are extremely large. There is no evidence of an exploitable vulnerability being fixed, and the commit message does not describe any security issue.

AI review queueditest: cover leases without expiryby yyforyongyu · 9dd64a3c · Oct 22, 2026 · 3 filesMessage 78 · AdequateInformational 15Details
Commit message · yyforyongyu

itest: cover leases without expiry

Verify reservation, restart, owner cancellation and first-spend cleanup
through WalletKit RPCs using a single input lifecycle.

Document the replacement flags, removed count response metadata and
explicit rejection of retired request parameters in the release notes.
Require boolean acknowledgement for callers migrating to no-expiry
leases.

78/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context✓ Mentions testing or verification
Why it was queued
signing or wallet pathsecond-pass: security-sensitive path
AI analysis · Informational 15/100

This commit only adds integration tests and release notes for a previously implemented feature: output leases that never expire. It does not change production code, so it cannot introduce a new security vulnerability or fix one directly.

Security candidatewalletrpc: reject retired lease requestsby yyforyongyu · e0269ab2 · Oct 22, 2026 · 2 filesMessage 68 · AdequateLow 48Details
Commit message · yyforyongyu

walletrpc: reject retired lease requests

Reject retired confirmation-count protobuf fields before lease
acquisition or PSBT funding so older clients cannot silently create
timed leases. Use preserved unknown fields and the existing protobuf
wire parser to keep the retired fields out of the public request schema.

Cover both public handlers with decoded old requests and stable error
identity assertions. Keep the existing wallet lifetime implementation.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
signing boundarysigning or wallet path
AI analysis · Low 48/100

This change is a hardening patch, not an active exploit fix. LND removed two protobuf fields that let callers request timed output leases. Older clients still send those field numbers over the wire as 'unknown' protobuf data, so the server would silently ignore them and create a different kind of lease than the client expected. The patch now detects those retired field numbers in LeaseOutput and FundPsbt and returns an explicit error instead of silently falling through. The main risk is operational: an old client or script that relied on the removed fields will now fail fast rather than silently changing behavior.