Daily Top · Threats / Cryptocurrency / Incident

LNbits said stop. v1.6.2 is not the patch.

Project, 8 November. Wave of attacks. Every operator told to shut the LNbits server down. Hosted my.lnbits.com taken offline. One operator reported 2,285,545 sats withdrawn through new accounts. No CVE. No GitHub release after September. Not Bitcoin. Not Lightning.

RWP Ventures · 2026-11-09 · event 2026-11-08 · 13 min read · priority 8.7

Bottom line up front

CONFIRMED On 8 November 2026 the LNbits project told every operator of its Lightning wallet and accounts software to stop their LNbits servers immediately after what it called a wave of attacks, with security patches still to ship. The hosted service at my.lnbits.com was taken down the same day. As of this run, GitHub’s latest release is still v1.6.2 (18 September 2026, a PostgreSQL backup hotfix). There is no CVE, no GitHub security advisory for this incident, and no KEV row.

Do not collapse five facts.

  1. This is the LNbits application, not Bitcoin and not Lightning. The stop-all covers LNbits servers. The project did not tell operators to shut Lightning nodes, on-chain wallets, or other wallet software. A chain reorg, a Lightning channel close, or a Core/LND bump does not close this ticket. The software layer that creates accounts and pays invoices is the asset.
  2. Paywall was the first warning, not a confirmed named bug. At about 12:33 UTC on 8 November, @lnbits asked operators with the Paywall extension enabled to disable it while the project looked at a possible bug. About 80 minutes later, at about 13:56 UTC, that instruction was superseded: all LNbits servers, not a subset. One operator later said a red-team review confirmed which extension was used on his instance. Neither he nor the project had named that extension as of The Hacker News’s 09:00 UTC 9 November cutoff. Do not ticket “disable Paywall and restart.”
  3. v1.6.2 is not the patch. The 1.7.0 release candidates are not the patch. Latest tagged stable release on GitHub this run is v1.6.2. Tags v1.7.0-rc1 / rc2 date to 29 October and are a version bump, not Sunday’s incident. Contributor Ben Arc estimated patches “should be merged and available within 24-48hrs” from his 8 November post — roughly 14:00 UTC 9 November to 14:00 UTC 10 November. That is an estimate, not a restart authorization. The Hacker News, writing 9 November, said that as of 09:00 UTC the project had not released a fixed version, had not cancelled the notice, and had not told operators how to check whether they were hit before they went dark. This run’s GitHub check agrees: no newer release.
  4. One operator’s 2,285,545 sats is not a census. Kevin Ravensberg (@KevinR4v), running BitTipBot, reported newly created accounts withdrew 2,285,545 satoshis (about 0.023 bitcoin) from his instance between 09:22 and 14:00 UTC on 8 November — starting more than three hours before the first public warning. He pledged to repay users. The project has not published a total, a victim count, or a version list. Treat any larger loss figure as UNKNOWN.
  5. This is not the September V12 bugs and not CVE-2025-32013. V12’s 29 September write-up is an authentication-cache bypass and a PayPal credit-before-capture pair, both fixed in v1.6.0. The 2025 LNURL-auth SSRF (CVE-2025-32013 / GHSA-qp8j-p87f-c8cc) is a different class. Do not close Sunday’s stop-all with a 1.6.0 or 1.0.0 bump you already applied.

The operator sentence: if you run LNbits, the server stays off until the project ships a release that is not v1.6.2 and tells you it is safe to come back. A Lightning node restart, a Paywall toggle, or a KEV row you do not have is not the close.

What happened

8 November 2026, Sunday, after yesterday’s 06:00 ET Daily Top. [Yesterday](/posts/2026-11-08-storm-3358-artex-financial/) was Microsoft’s Storm-3358 / ARTEX financial-sector note. This ticket opened in the afternoon UTC.

High-level timeline. Defensive only. Not a reproduction.

Time (UTC, 8–9 Nov)What is confirmed
09:22 8 NovRavensberg’s instance: attack window starts (his report)
12:33 8 Nov@lnbits: operators with Paywall enabled, disable it immediately while a possible bug is investigated
13:56 8 Nov@lnbits: all LNbits users, stop servers immediately; “a number of security patches to push”
~14:00 8 NovRavensberg’s reported attack window ends
Later 8 Nov@lnbits: all instances on hosted my.lnbits.com disabled as a precaution; services “within 24 hours” of that notice (The Hacker News counted that as about 17:20 UTC 9 November)
09:00 9 NovThe Hacker News cutoff: no fixed version, notice not cancelled, no self-check procedure published
This run, 9 NovGitHub lnbits/lnbits latest release still v1.6.2 (18 September). Catalog still 2026.11.06 / 1769

The shutdown notice covers LNbits servers. It did not exempt a funding-source configuration. It did not name versions. It did not name the bug class beyond “a wave of attacks” and, in secondary reporting of operator accounts, new accounts withdrawing funds they should not have been able to touch.

Do not publish a procedure for opening those accounts or moving the sats. Defenders need the stop, the scope, and the non-equivalence with Lightning.

Why it matters

Three operational reasons this is the daily.

First: the vendor instruction is a full stop, not a patch-when-convenient advisory. That is rare. It arrived after yesterday’s cron, with no CVE and no GitHub tag to point a change ticket at. Shops that wait for KEV, NVD, or a SemVer bump will keep a payment front-end on the internet through an active drain window.

Second: the false close is “we are on Lightning” or “we disabled Paywall.” LNbits sits on top of a Lightning node as a multi-wallet accounts layer — tips, paywalls, SaaS instances, community custodial fronts. The node can be healthy while the accounts layer is writing withdrawals. The first public warning named Paywall; the binding instruction is all servers. Restarting into v1.6.2 because “that was the last hotfix” reopens the door the project just closed.

Third: there is still no KEV interrupt overnight, so this is the increment. Live catalog remains 2026.11.06 / 1769. [SonicWall CVE-2026-19868](/posts/2026-11-07-sonicwall-cve-2026-19868-kev/) is still due 20 November. Brotli CVE-2026-111497 and Adobe Commerce CVE-2026-48356 are due tomorrow, 10 November. None of those is an LNbits process. Waiting for a catalog row before taking a Lightning accounts server offline is how the next withdrawal lands.

This is not confirmed OT or ICS process impact. A plant or utility that used LNbits for a community wallet still has an IT payment-app incident until someone shows the same host reaches a controller network.

Who / what is affected

Confirmed in scope: anyone running an LNbits server — self-hosted or on the project’s hosted my.lnbits.com — as of 8 November, until the project ships a post-incident release and lifts the stop.

Confirmed loss, one operator: BitTipBot / Kevin Ravensberg, 2,285,545 sats, new accounts, 8 November 09:22–14:00 UTC. He said users cannot reach funds until LNbits patches. That is one instance, not a global total.

Confirmed not this ticket:

UNKNOWN total sats moved; number of instances hit; whether Paywall is the exploited extension; affected version range; whether v1.7.0-rc builds are safer or worse; whether hosted my.lnbits.com actually returned at the 24-hour mark.

Technical context

High level. This is not a how-to.

LNbits is an open-source Lightning wallet and accounts system. Operators run it as a web service in front of a Lightning backend. It creates wallets, issues invoices, and pays. Extensions add features (Paywall was the one named in the first warning). The attack reports describe account creation plus withdrawal on a live instance. That is enough to treat the HTTP surface of LNbits as the door. It is not enough to name a CVE, a route, or an extension.

Assessed ATT&CK from the public reporting, not from a vendor navigator: T1190 Exploit Public-Facing Application (if the bug is in the web app — ASSESSED, CVE unnamed); T1078 Valid Accounts once a new account exists; T1657 Financial Theft for the withdrawals. Do not invent T1048 exfil or a ransomware overlay. None was reported.

September is background, not this bug. V12 found two high-impact issues in August, both first shipping in v1.4.0, both fixed in v1.6.0 (2 September): a scoped API token becoming wallet-admin through an auth-cache miss, and PayPal balance credited before capture. Those are closed in 1.6.0. Sunday’s stop-all landed on a tree whose latest release is 1.6.2. Either a new bug, a missed extension, or a regression. The project has not said which. Do not brief “we patched in September.”

Threat / exploitation status

ClaimStatusConfidence
LNbits told all operators to stop servers immediately on 8 November after a wave of attacksCONFIRMEDHigh (project X; THN 9 Nov)
Hosted my.lnbits.com instances disabled the same dayCONFIRMEDHigh
Paywall disable was the first warning and was superseded by the all-server stopCONFIRMEDHigh
Named CVE / GitHub advisory / KEV row for this incidentNONE this runHigh (GitHub releases + catalog 2026.11.06 / 1769)
Latest GitHub release still v1.6.2 (18 Sep)CONFIRMED this runHigh
2,285,545 sats withdrawn from one BitTipBot instance via new accountsREPORTEDModerate-high (operator first-person; not a project total)
Exploited extension namedUNKNOWNProject and operator had not named it as of 09:00 UTC 9 Nov
V12 Sep / CVE-2025-32013 is this bugCONFIRMED false as publishedHigh (different issues, already patched)
Bitcoin L1 or Lightning protocol compromisedCONFIRMED false as publishedHigh (software-layer incident)
OT / ICS process impactUNKNOWNDo not infer
Restart authorizationNONEHigh (stop still in force at THN’s 09:00 UTC 9 Nov cutoff; GitHub unchanged this run)

What defenders should do

  1. If you run LNbits, leave it off. Self-hosted and hosted. Do not restart into v1.6.2. Do not treat v1.7.0-rc1/rc2 (29 October) as Sunday’s fix. Wait for a project release and an explicit lift of the stop. If you already stopped, keep it stopped.
  2. Inventory what sat behind that process. Which Lightning node, which on-chain hot wallet, which fiat rails, which Telegram/bots (BitTipBot-class), which customer balances. The node may still be up. Treat custodial balances on that LNbits instance as possibly moved until you can read the books on a patched build. Do not move remaining funds through the same LNbits HTTP API while it is unpatched.
  3. Do not close the ticket as “Paywall disabled.” The binding instruction is all servers. Hunt new accounts, unexpected withdrawals, and extension state after you have a patched build to stand up in a controlled window — not by turning the vulnerable server back on to look.
  4. Do not merge this with yesterday’s ARTEX hunt or with BOSSAI. Storm-3358 is anonymous upload/SSRF then WISP/vshell/AdaptixC2 on financial web apps. BOSSAI is a Taiwan-based operator’s 12-agent mesh against emergency APIs (Acronis 4 November; THN recap 9 November). Different IOCs, different assets. An LNbits shutdown does not close a bank’s upload form.
  5. Keep the federal clocks that are actually on the catalog. Brotli CVE-2026-111497 and Adobe Commerce CVE-2026-48356 due 10 November. ESX CVE-2026-41709 due 11 November. Ivanti CVE-2026-118527 due 13 November. Apple CVE-2024-27834 due 18 November. SonicWall CVE-2026-19868 due 20 November. OpenSSH CVE-2026-123158 still not KEV. UnrealIRCd CVE-2026-114713 is not KEV; if you run linked UnrealIRCd 6, that is a separate ticket (6.2.7.1 or hot-patch bundle-20261107; maintainer said 8 November that exploitation was being seen).
  6. OT / ICS operators: only if LNbits is actually in your environment (community wallet, visitor payment, contractor tips). Then it is an IT payment-app outage plus a possible theft, not a setpoint change. Do not record process impact you do not have.

RWP assessment

This is the daily because it is a vendor stop-all after confirmed theft, published after yesterday’s 06:00 ET Storm-3358 run, with no patch to point at, and it is not Bitcoin, not Lightning, not last month’s V12 bugs, and not a KEV reprint. Evidence quality is high on the project’s instruction and on the absence of a GitHub release. Evidence quality is operator-first-person on the 2.29 million sat figure — we label that REPORTED, one instance. Evidence quality is none on the CVE, the extension name, the total, and OT impact. We are not naming Paywall as the exploited component. We are not publishing account-creation or withdrawal steps. We are not treating v1.7.0-rc as a security release.

Score 8.7. It outranks UnrealIRCd CVE-2026-114713 (vendor confirmed exploitation 8 November, but a patch and a SPAMFILTER workaround exist; IRC, not a payment front-end with a stop-all), Acronis BOSSAI (strong CI/emergency-services research, 4 November primary, THN recap today — not a 24-hour increment), OpenSSH CVE-2026-123158 (still not KEV; already the 7–8 November runner-up), WP Duplicate CVE-2026-123751 (WPScan exploited-in-the-wild, no fix, but WordPress.org 200+ installs), Crédit Mutuel Arkéa (already 8 November candidate), Elastic’s eight-minute infostealer recap (Help Net Security 9 November of a 4–5 November report), and the 10–20 November KEV clocks.

What we are watching

Sources

Sources

  1. The Hacker News — LNbits Urges All Operators to Shut Down Their Bitcoin Lightning Wallet Servers Amid Attacks, 9 Nov 2026
  2. LNbits — Stop-all alert on X, 8 Nov 2026
  3. LNbits — Paywall precaution on X, 8 Nov 2026
  4. LNbits — Hosted service disabled on X, 8 Nov 2026
  5. Ben Arc — Patch estimate on X, 8 Nov 2026
  6. Kevin Ravensberg — Loss report on X, 8 Nov 2026
  7. LNbits GitHub — Releases (v1.6.2 latest as of this run)
  8. V12 — Two high-impact vulnerabilities in LNbits, 29 Sep 2026
  9. CISA — KEV JSON feed catalog 2026.11.06
  10. CISA — Known Exploited Vulnerabilities Catalog
  11. UnrealIRCd — oss-sec CVE-2026-114713, 8 Nov 2026
  12. NVD — CVE-2026-114713
  13. Acronis TRU — BOSSAI, 4 Nov 2026
  14. RWP Ventures — Storm-3358 ARTEX Daily Top 2026-11-08
  15. RWP Ventures — SonicWall CVE-2026-19868 KEV Daily Top 2026-11-07
  16. RWP Ventures — CyberStrikeAI LLM post-exploitation Daily Top 2026-11-02