LL
← All projectsLightning Labs

LND

Go implementation of a complete Lightning Network node.

BitcoinLightning NetworkNormal
Repository coverage

1241 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.

118security candidates168second-pass queue1236AI analyses
55commits · 30 days
171commits · 60 days
830commits · 180 days
1219commits · 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.

67/100 average clarity
279Strong · 80–100
564Adequate · 60–79
367Thin · 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 Osuntokun19940198474
Elle Mouton26832268167
ziggie41015408164
Jared Tobin909171
Erick Cestari26726070
bitromortac46546067
yyforyongyu55355065
saubyk15215070
Oli727057
Suheb323075
Yong323073
elnosh323071
Analysis record

Published AI watches

Last scanned 56 minutes ago

Moderate 59 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11023 from ellemouton/update-fee-log

This commit hardens how LND handles Lightning channel fee updates and mailbox message admission. It enforces the protocol rule that only the channel opener can send fee updates, prevents uncommitted fee updates from piling up in the update…

BOLT 2 role validation for UpdateFee senderReplacement of uncommitted fee updates to prevent log bloat and index gapsBounded mailbox wire-message queue (count and bytes)
7f56541dby Yong+750−138 files
No security note in commit
Moderate 57 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11019 from Roasbeef/coop-close-race-public

This update fixes two related problems in LND's cooperative channel-closing code. First, it eliminates a data race where the same state machine was being advanced by two different internal goroutines at once, which could cause crashes or i…

Data race in legacy cooperative close state machine driven concurrently by link and peer goroutinesMissing or inconsistent validation of remote delivery script when no upfront shutdown script is recordedpanic() in DustLimitForSize on unrecognized script sizes replaced with safe fallback
f4ae5656by Yong+463−609 files
Vendor flagged security relevance
Moderate 63 AI analysisMessage 58 · Thin
LL Lightning LabsLND BitcoinLightning Network

Merge pull request #11024 from ziggie1984/invoices-update-validation

This commit tightens how LND processes invoice payments, especially for newer multi-path (MPP) and AMP invoices, keysend payments, and replayed payments. It adds checks that ensure the right payment preimage is used for each invoice type a…

Preimage validation added for AMP and regular invoice replaysAMP invoices forced into MPP update path, preventing legacy-path processingLegacy settlement now fails when invoice-level preimage is missing
61f3b5a7by ziggieXXX+482−114 files
No security note in commit
Moderate 59 AI analysisMessage 35 · Opaque
LL Lightning LabsLND BitcoinLightning Network

invoices: refine update handling

This commit tightens how LND handles invoice payment updates, especially for newer 'AMP' invoices and old-style 'keysend' payments. It adds checks so that the wrong kind of payment cannot be processed against the wrong kind of invoice, and…

Added preimage/hash mismatch checks for both regular and AMP invoice replaysAMP records now require an MPP payload, preventing AMP processing on the legacy pathLegacy path now rejects AMP invoices and invoices missing invoice-level preimages
6be6350eby ziggie+476−113 files
No security note in commit
Informational 15 AI analysisMessage 57 · Thin
LL Lightning LabsLND BitcoinLightning Network

docs: update 0.21.2 release notes

This commit only updates the release notes document for version 0.21.2. It adds a bullet point describing a previously merged code change about invoice handling and adds a contributor name to the credits list. No actual code was changed, s…

758bbb8eby ziggie+6−01 file
No security note in commit
Moderate 60 AI analysisMessage 88 · Strong
LL Lightning LabsLND BitcoinLightning Network

peer+lnwallet/chancloser: advance the legacy closer from one goroutine

This commit fixes a concurrency bug in how LND negotiates cooperative channel closures. Previously, two different goroutines could update the same channel-closing state machine at the same time, which could corrupt internal data or cause a…

Data race on ChanCloser state field detected under go test -raceConcurrent access to priorFeeOffers map and signing step by two goroutinesSingle-goroutine ownership invariant now documented on ChanCloser type
e5e134ddby Olaoluwa Osuntokun+292−384 files
Vendor flagged security relevance
Low 29 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

lnwallet/chancloser: record the remote close output only when accepted

This change fixes a bookkeeping bug in how LND records the other party's preferred closing address during a channel close. Previously, a late or unexpected Shutdown message could overwrite the correct address with one from a message that w…

State variable written before validationRejected message could overwrite accepted close outputFix prevents stale/incorrect close metadata
fb89732dby Olaoluwa Osuntokun+7−21 file
No security note in commit
High 72 AI analysisMessage 68 · Adequate
LL Lightning LabsLND BitcoinLightning Network

htlcswitch: bound peer-controlled channel ingress

This commit hardens LND's channel message handling against a peer that tries to overwhelm the node. It adds limits on how many ordered messages and how much data a peer can queue up for a channel, and it disconnects the peer if either limi…

Adds resource bounds on peer-controlled channel ingress (count + bytes)Disconnects peer on mailbox admission failure instead of silently dropping ordered channel messagesPrevents unauthorized UpdateFee from channel initiator before fee-exposure calculations
ef24f2c5by Elle Mouton+507−94 files
Vendor flagged security relevance
Moderate 64 AI analysisMessage 83 · Strong
LL Lightning LabsLND BitcoinLightning Network

lnwallet/chancloser: validate delivery scripts in the RBF closer

This commit fixes a bug in LND's cooperative channel-closing code where a peer could supply an invalid, empty, or even money-burning Bitcoin address during a close, and LND would accept it without checking. The fix makes the newer 'RBF' cl…

Missing input validation on remote-supplied shutdown/delivery scriptEmpty/zero-length script accepted as valid due to nil-treatment in validatorMid-negotiation script swap via ClosingComplete not validated
a8e2a0f7by Olaoluwa Osuntokun+131−142 files
Vendor flagged security relevance
Moderate 51 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

lnwallet: make DustLimitForSize total over the sizes it can be handed

This change removes a deliberate program crash (panic) in LND's fee/dust-limit helper and replaces it with a safe fallback. Previously, if the helper received a script length it didn't explicitly recognize, it would panic and potentially c…

Removal of a panic/default branch in favor of a conservative fallbackDefensive handling of arbitrary and future witness-version script sizesPotential denial-of-service vector eliminated: unhandled script lengths no longer crash the helper
f80f92dcby Olaoluwa Osuntokun+23−62 files
No security note in commit
Informational 15 AI analysisMessage 57 · Thin
LL Lightning LabsLND BitcoinLightning Network

docs: move gossip bounds note to v0.21.2

This commit only moves an existing release-note entry from one version's notes (v0.22.0) to another (v0.21.2). It does not change any code, build files, configuration, or tests. The change is purely editorial documentation bookkeeping.

83e7eb8cby ziggie+11−112 files
No security note in commit
High 72 AI analysisMessage 68 · Adequate
LL Lightning LabsLND BitcoinLightning Network

discovery: bound channel range reply buffering

This change fixes a way a malicious or buggy peer could make an LND node use an unpredictable amount of memory while syncing the Lightning channel graph. Before the fix, the node only limited the size of one compressed reply message, but a…

Resource exhaustion / unbounded memory growth in gossip syncRemote peer can influence memory allocation via ReplyChannelRange streamMissing aggregate limit on decoded working set
ceff94faby Olaoluwa Osuntokun+335−153 files
Vendor flagged security relevance
Low 33 AI analysisMessage 47 · Thin
LL Lightning LabsLND BitcoinLightning Network

docs: update release notes

This commit updates the release notes for LND 0.21.2 to describe a fix for a bug in the HTLC interceptor RPC. The bug caused LND to fail to forward payments in blinded routes when the next hop was identified by node ID instead of channel I…

Blinded route forwarding failure could disrupt payment routing or be used to probe path structure.Sentinel value change in RPC API could cause client misclassification of forwards if not handled.Release notes describe the change as a fix for a forwarding issue, not as a security vulnerability.
f42b4298by bitromortac+20−01 file
No security note in commit
Informational 14 AI analysisMessage 83 · Strong
LL Lightning LabsLND BitcoinLightning Network

itest: cover blinded route next_node_id forwarding

This commit only adds new integration tests for an existing feature in LND: forwarding blinded payments where the next hop is identified by node ID rather than channel ID. It does not change production code, fix a bug, or introduce any new…

da6a40c0by bitromortac+432−02 files
No security note in commit
Low 25 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

witness beacon: report node-ID next hop to the on-chain HTLC interceptor

This change fixes a reporting gap in LND's 'witness beacon,' a component that watches on-chain transactions and tells the HTLC interceptor where a payment should go next. Previously, when the next hop was identified by a node ID rather tha…

Missing security-critical metadata in on-chain interception pathBehavioral parity between on-chain and off-chain HTLC interceptionPotential misclassification of node-ID forwards as final receives
9c4b8bfeby bitromortac+59−52 files
No security note in commit
Moderate 59 AI analysisMessage 81 · Strong
LL Lightning LabsLND BitcoinLightning Network

htlcswitch: forward node-ID blinded hops via non-strict forwarding

This commit fixes a bug in LND's payment forwarding for 'blinded routes'—a privacy feature in the Lightning Network. Previously, if the next hop was identified by node ID rather than a specific channel ID, the payment could not be forwarde…

Fixes a functional forwarding failure for blinded-route payments using node-ID next hops (issue #10937).Prevents private channel SCID leakage in failure messages for node-ID blinded hops by returning FailUnknownNextPeer instead of a channel_update.Adds circular-route filtering for node-ID next hops before non-strict forwarding selection.
dbc57040by bitromortac+334−416 files
No security note in commit
Informational 19 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

htlcswitch+lnrpc: report node-ID next hop to the off-chain HTLC interceptor

This change improves how the Lightning Network Daemon (LND) reports forwarding details to external plugins that inspect HTLCs (payment packets). For a new type of blinded route where the next hop is identified by node ID rather than a spec…

Information-correctness fix for HTLC interceptor APIPrevents misclassification of node-ID forwards as exit hopsAdds explicit sentinel to disambiguate zero SCID semantics
32373b76by bitromortac+29−13 files
No security note in commit
Low 25 AI analysisMessage 73 · Adequate
LL Lightning LabsLND BitcoinLightning Network

htlcswitch: classify a node-ID forward as a forward event

This commit fixes a bookkeeping bug in LND's HTLC event stream. When a payment is forwarded to a 'blinded' next hop identified only by a node public key (not a channel ID), and it fails before the switch picks an outgoing channel, the even…

Event misclassification in HTLC event stream (forward reported as receive)Blinded/node-ID forwarding path introduced prior to this fixFailure and resolution packets dropped next-hop metadata before fix
a4844ef5by bitromortac+149−04 files
No security note in commit
Low 29 AI analysisMessage 68 · Adequate
LL Lightning LabsLND BitcoinLightning Network

htlcswitch/hop: decode next_node_id blinded hops

This commit adds support in LND for a different way of identifying the next hop in a private ('blinded') Lightning payment route. Previously LND expected the next hop to be identified by a channel ID; now it also accepts the next node's pu…

New validation enforces BOLT 4 mutual exclusivity of short_channel_id and next_node_id in blinded route dataPreviously ambiguous or unhandled next_node_id forwarding case now explicitly decodedNo memory-safety, cryptographic, or authorization changes observed
4fd4289aby bitromortac+471−86 files
No security note in commit
Moderate 60 AI analysisMessage 78 · Adequate
LL Lightning LabsLND BitcoinLightning Network

lnwire: cap decoded short channel IDs

This commit tightens a safety limit in LND's Lightning wire protocol parser. Previously, a compressed network message could decompress into an unbounded number of short channel IDs, potentially consuming far more memory than the wire messa…

Resource-limit bypass: compressed-input bound did not bound decoded outputMemory-exhaustion risk from zlib decompression of peer-supplied dataTruncated/corrupt zlib streams previously accepted as valid partial results
d1622919by Olaoluwa Osuntokun+271−502 files
No security note in commit
Repository ledger

Explore captured commits

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

Lower-prioritygraph/db: use batch fetching for DeleteChannelEdgesby Elle Mouton · 556af8e2 · Aug 7, 2025 · 2 filesMessage 50 · ThinInformational 10Details
Commit message · Elle Mouton

graph/db: use batch fetching for DeleteChannelEdges

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Informational 10/100

This commit refactors how LND deletes channel edges in its graph database so that it fetches related channel data in batches rather than one by one. The change is a performance optimization and code cleanup. There is no direct evidence in the commit or supplied references that this fixes a security vulnerability.

Lower-prioritygraph/db: refactor buildNode to not take a pointerby Elle Mouton · ce3401ee · Aug 7, 2025 · 2 filesMessage 65 · AdequateInformational 15Details
Commit message · Elle Mouton

graph/db: refactor buildNode to not take a pointer

Since the type returned from the DB is not a pointer. This will be
useful later on.

65/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body
AI analysis · Informational 15/100

This commit is a small internal code cleanup in LND's graph database code. It changes a helper function so it receives a copy of a data record instead of a pointer to that record. There is no user-facing change, no bug fix, and no security relevance visible in the diff or commit message.

Lower-prioritygraph/db: refactor makeZombiePubkeysby Elle Mouton · f51adaf3 · Aug 7, 2025 · 2 filesMessage 70 · AdequateInformational 15Details
Commit message · Elle Mouton

graph/db: refactor makeZombiePubkeys

Let the helper method only take the params it needs so that we dont need
to construct an entire models.ChannelEdgeInfo object to pass to it. This
will be useful later on.

70/100 · AdequateMessage clarity
✓ Descriptive subject✓ Provides detailed explanatory context✓ Explains rationale or failure mode
AI analysis · Informational 15/100

This commit is a straightforward code cleanup that changes a helper function to accept individual values instead of a whole object. It does not appear to fix or introduce any security issue. However, the diff contains one suspicious typo-like change in the default case where the function now returns `node1` instead of `node2`, which looks like a potential bug but is not a security vulnerability on its own.

Lower-prioritygraph/db: add ChanUpdatesInHorizon benchmarkby Elle Mouton · 4cc5cfef · Aug 7, 2025 · 1 fileMessage 45 · ThinInformational 15Details
Commit message · Elle Mouton

graph/db: add ChanUpdatesInHorizon benchmark

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Informational 15/100

This commit only adds a new benchmark test that measures how fast a database method called ChanUpdatesInHorizon runs. It does not change any production code, fix a bug, or alter behavior. There is no security relevance.

Lower-prioritydocs: move v0.19.3 items to correct fileby Oliver Gugger · 429aa830 · Aug 6, 2025 · 2 filesMessage 72 · AdequateInformational 15Details
Commit message · Oliver Gugger

docs: move v0.19.3 items to correct file

Moves the PRs that are going to be shipped with v0.19.3 into the correct
release notes file.

72/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope✓ Provides an explanatory body
Why it was queued
documentation-only discount
AI analysis · Informational 15/100

This commit is purely a documentation housekeeping change. It moves release note entries for three already-merged changes (a logging fix, a deadlock fix, and gossip bandwidth default increases) from the v0.20.0 file to the v0.19.3 file because they are actually shipping in the earlier v0.19.3 release. No code behavior is changed.

Security candidatescripts: update ffranr release signing keyby ffranr · 0676c566 · Aug 6, 2025 · 2 filesMessage 68 · AdequateInformational 21Details
Commit message · ffranr

scripts: update ffranr release signing key

Updated the release signing key. New key fingerprint:

C97A AA14 70F9 7987 8F7A 6DED C344 0ACF 100A 33B4

The old key is no longer in use.

68/100 · AdequateMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
Why it was queued
signing boundary
AI analysis · Informational 21/100

This commit swaps out one PGP public key for another in LND's release verification script. The old key belonged to the same person (ffranr/Robyn Ffrancon) and is no longer used; the new key is the current release signing key. There is no code vulnerability here—it's a routine key rotation. However, if someone could silently replace the key with an attacker's key, they could trick users into trusting fake LND releases. The commit itself appears legitimate and matches the stated purpose.

Lower-priorityscripts: remove expired pgp keyby ziggie · 3afac744 · Aug 6, 2025 · 2 filesMessage 45 · ThinInformational 19Details
Commit message · ziggie

scripts: remove expired pgp key

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Informational 19/100

This commit removes an expired PGP public key from the install verification script. It is a routine cleanup: keeping an expired key in the script would not let attackers take over systems, but it could cause install verification to fail or confuse users. There is no indication this is a security vulnerability fix.

Lower-prioritymulti: remove the NodeRTx interface typeby Elle Mouton · 8aa8c7cc · Aug 6, 2025 · 10 filesMessage 45 · ThinInformational 15Details
Commit message · Elle Mouton

multi: remove the NodeRTx interface type

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Informational 15/100

This commit is a straightforward internal code cleanup in LND's channel graph code. It removes an unnecessary intermediate interface called NodeRTx and passes the underlying LightningNode object directly to callers. There is no security-relevant change here—no bug is fixed, no vulnerability is introduced, and no behavior visible to users changes.

Lower-prioritygraph/db: remove NodeRTx ForEachChannel methodby Elle Mouton · 15e17e12 · Aug 6, 2025 · 4 filesMessage 45 · ThinInformational 12Details
Commit message · Elle Mouton

graph/db: remove NodeRTx ForEachChannel method

45/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Informational 12/100

This commit removes an unused method called ForEachChannel from the graph database interfaces and both of its implementations (one for the older key-value store and one for the newer SQL store). It also updates one integration test to use a different, already-existing method that provides the same information in a cached form. There is no indication this fixes a security bug; it appears to be routine code cleanup to simplify the interface and remove duplicated functionality.

Lower-priorityautopilot: remove the ForEachChannel method from Node interfaceby Elle Mouton · 699e3359 · Aug 6, 2025 · 5 filesMessage 73 · AdequateInformational 15Details
Commit message · Elle Mouton

autopilot: remove the ForEachChannel method from Node interface

And instead make use of the new ForEachNodesChannels method which
uses a much more efficient method for iterating through nodes&channels.

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
AI analysis · Informational 15/100

This commit is a routine internal refactoring of LND's autopilot (automatic channel management) code. It removes an older, less efficient method for looping through channels and replaces it with a newer, more efficient one. There is no indication this fixes a security bug or introduces a vulnerability.

Lower-priorityautopilot: add ForEachNodesChannels method to ChannelGraph interfaceby Elle Mouton · ce7fe84d · Aug 6, 2025 · 3 filesMessage 65 · AdequateInformational 15Details
Commit message · Elle Mouton

autopilot: add ForEachNodesChannels method to ChannelGraph interface

This is in preparation for removing the ForEachChannel method from the
Node interface.

65/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides an explanatory body
AI analysis · Informational 15/100

This commit is a routine internal refactoring in LND's autopilot (automated channel-opening) module. It adds a new method called ForEachNodesChannels to the ChannelGraph interface and implements it for database-backed and cached graph types, plus an in-memory test mock. There is no user-facing behavior change, no bug fix, and no security-sensitive logic. It simply prepares the codebase for a later removal of an older ForEachChannel method on the Node interface.

Lower-prioritygraph/db: let ForEachNodeCached maybe fetch node addressesby Elle Mouton · 5727bfa6 · Aug 6, 2025 · 9 filesMessage 73 · AdequateInformational 15Details
Commit message · Elle Mouton

graph/db: let ForEachNodeCached maybe fetch node addresses

Here we adjust the ForEachNodeCached graph DB method to pass in a node's
addresses into the provided call-back if requested. This will allow us
to improve the performance of node/channel iteration in the autopilot
subserver.

73/100 · AdequateMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component✓ Provides detailed explanatory context
AI analysis · Informational 15/100

This commit is a routine performance and API improvement. It extends an internal graph database method so callers can optionally ask for each node's network addresses while iterating. When addresses are not needed, the code skips an extra database lookup, which makes operations like autopilot faster. There is no security fix or vulnerability here.

Lower-priorityitest: add payment test with max htlc restrictionby ziggie · 5c22e5eb · Aug 6, 2025 · 2 filesMessage 55 · ThinInformational 12Details
Commit message · ziggie

itest: add payment test with max htlc restriction

55/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Mentions testing or verification! No meaningful explanatory body
AI analysis · Informational 12/100

This commit only adds a new integration test to the LND project. It checks that when a Lightning channel has a maximum HTLC (payment chunk) size set, the software correctly splits a larger payment into smaller pieces that fit within that limit. There is no code fix or behavior change in the main application—only a test was added.

Lower-prioritydocs: fix 19.2 release-notesby ziggie · 28f286a2 · Aug 6, 2025 · 1 fileMessage 57 · ThinInformational 15Details
Commit message · ziggie

docs: fix 19.2 release-notes

57/100 · ThinMessage clarity
✓ Descriptive subject✓ Names a concrete action or component✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
documentation-only discount
AI analysis · Informational 15/100

This commit only updates the release notes document by adding missing names to the contributors list. It does not change any code, configuration, or security behavior.

Lower-prioritydocs: add release-notesby ziggie · a19e7f2e · Aug 6, 2025 · 1 fileMessage 40 · ThinInformational 15Details
Commit message · ziggie

docs: add release-notes

40/100 · ThinMessage clarity
✓ Subject identifies a change✓ Uses a recognizable type or scope! No meaningful explanatory body
Why it was queued
documentation-only discount
AI analysis · Informational 15/100

This commit only adds a release note describing a previously fixed bug. It does not change any code, cryptographic logic, network behavior, or configuration. There is no security issue introduced or addressed by this documentation-only change.

Lower-prioritymulti: skip range check in pathfinder and switch for custom htlc paymentsby ziggie · e2a9b172 · Aug 6, 2025 · 4 filesMessage 50 · ThinLow 35Details
Commit message · ziggie

multi: skip range check in pathfinder and switch for custom htlc payments

50/100 · ThinMessage clarity
✓ Specific, descriptive subject✓ Names a concrete action or component! No meaningful explanatory body
AI analysis · Low 35/100

This commit changes how the Lightning Network Daemon (LND) handles 'custom HTLC' payments. Normally, LND enforces minimum and maximum amount limits on forwarded payments to prevent abuse and routing loops. For custom HTLCs—special payments that carry extra data and may use non-Bitcoin assets—these amount checks are now skipped in both the routing pathfinder and the actual forwarding switch. The change is described by the authors as intentional and needed to avoid routing loops, not as a security fix. It narrows a safety guard for a specific payment type, which could matter if the custom-channel logic has bugs or is misused.