fix(content): never take a paid buyer's ecash and then fail to deliver

2026-09-18: a peer purchase paid 10 sats, the seller redeemed them, and the
buyer got no file plus a "seller doesn't accept your Cashu mint" error.
Three defects lined up:

1. The seller checked file existence with stat() but only read the file
   AFTER redeeming the payment. Filebrowser-owned 0640 files (uid 100999)
   passed stat but failed fs::read for the archipelago service user.
   serve_content now checks existence and readability BEFORE the payment
   gate, so an unservable file costs the buyer nothing.
2. The HTTP handler mapped every serve_content error to a bare, unlogged
   404. A server-side failure is now a logged 500. (A 404 also makes the
   buyer's Auto transport re-send the request over Tor.)
3. That re-send carried the same single-use token, which the mint had
   already spent, so the seller answered 402. Redemption is now
   idempotent: a token that verified for an item keeps authorising that
   item for 10 minutes (per token, per item; SHA-256 keyed, in-memory,
   concurrent requests serialised, failures never cached).

Buyer side: reclaim_spent_ecash now reports whether the refund worked, and
the error text no longer claims "refunded" when it wasn't, or asserts the
seller rejects the mint when the cause is unknown.

Adds tests for replay, concurrency, failure-not-cached, cross-item, and
unreadable-file-before-payment.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

AGENTS.md

@@ -0,0 +1,23 @@
+# Blocking incident: Framework LND startup and false zero balance
+
+Before starting or resuming work in this repository, read
+`docs/incident-framework-lnd-startup.md` and check its status.
+
+The user explicitly required this incident to take priority over later work.
+While its status is OPEN:
+
+- Surface this blocker at session start, including when the user asks for unrelated work.
+- Prioritize investigation and repair on the actual Framework node. If access is
+  unavailable, ask for it and continue useful offline investigation; defer unrelated
+  work unless the user explicitly overrides this priority.
+- A manual LND restart, a source patch, passing local tests, or publishing an OTA
+  does not resolve the incident. Do not mark it fixed until the Framework's startup,
+  Receive flow, and balance behavior are verified on the node, including a controlled
+  reboot with access and recovery arrangements in place.
+- Preserve wallet identity, wallet/channel databases, credentials, and backups.
+  Never run wallet wipe/recreation as an automatic investigation or recovery step.
+- Record evidence, changes, validation, and remaining work in the incident document.
+
+This priority comes from the user's explicit instruction on 2026-09-15. It remains
+in effect across sessions until the documented acceptance criteria are met or the
+user explicitly changes it.

CHANGELOG.md

@@ -2,6 +2,26 @@
 
 ## Unreleased
 
+## v1.8.17-alpha (2026-09-15)
+
+- Minibits claims that every mint reports as already spent leave the retry queue, clearing repeated failure notices. Network errors and mixed mint failures remain queued for another attempt.
+- Minibits polls its primary relay first and connects to public fallback relays only when the primary is unreachable, reducing unnecessary connections.
+- Large payment backlogs are fetched from newest to oldest with a saved cursor, so polling can resume after interruptions or page limits. Payments sharing the same timestamp remain reachable.
+- Added regression coverage for spent-claim classification, wrapped and mixed mint errors, same-second payments, and interrupted or multi-poll backlogs.
+
+## v1.8.16-alpha (2026-09-15)
+
+- App updates refresh and verify the signed catalog before changing containers. A failed refresh or manifest reload cancels the update, and automatic updates wait for a successful refresh.
+- Fixed repeated Mempool update offers: downstream `-archyN` patches now sort above their upstream release, and moving a published image between registry namespaces does not hide a genuine upgrade.
+- Updates inspect installed component versions, refuse known downgrades, skip containers already at the target versions, and verify the resulting versions before reporting success.
+- Added regression coverage for stale catalogs, matching versions, publisher namespace changes, stack component updates, and keeping running containers untouched when no upgrade is needed.
+
+## v1.8.15-alpha (2026-09-13)
+
+- Cuprate is presented as one user-facing app in My Apps, including its UI launch button; the generated dashboard companion is hidden as an implementation detail instead of appearing under Services.
+- Added regression coverage for Cuprate install and installed-state grouping.
+- Release validation was rerun on the corrected tree before OTA and ISO publication.
+
 ## v1.8.14-alpha (2026-09-13)
 
 - **Cuprate gains a first-party companion dashboard.** The Monero node now has a Bitcoin-style status UI, safe app grouping, a 450 GB disk-safety gate, and a restricted RPC that is never exposed as a launch page.

app-catalog/catalog.json

@@ -378,13 +378,13 @@
     {
       "id": "mempool",
       "title": "Mempool Explorer",
-      "version": "3.0.0",
+      "version": "3.3.1-archy1",
       "description": "Bitcoin mempool and blockchain explorer. Real-time transaction and block visualization.",
       "icon": "/assets/img/app-icons/mempool.webp",
       "author": "Mempool",
       "category": "money",
       "tier": "core",
-      "dockerImage": "source.archipelago-foundation.org/lfg2025/mempool-frontend:v3.3.1",
+      "dockerImage": "source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1",
       "repoUrl": "https://github.com/mempool/mempool",
       "requires": [
         "bitcoin-knots",

apps/archy-mempool-web/manifest.yml

@@ -1,7 +1,7 @@
 app:
   id: archy-mempool-web
   name: Mempool Web
-  version: 3.0.1
+  version: 3.3.1-archy1
   # Where this app comes from, so scripts/check-upstream-releases.py can
   # tell us when the pin below has fallen behind. Without it nothing can:
   # container.image names our mirror, not the project it was mirrored from.
@@ -12,7 +12,7 @@ app:
   container_name: mempool
 
   container:
-    image: source.archipelago-foundation.org/lfg2025/mempool-frontend:v3.3.1
+    image: source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1
     pull_policy: if-not-present
     network: archy-net
 
@@ -45,7 +45,9 @@ app:
     # first, but nginx binds 0.0.0.0:8080 (IPv4) only -> localhost probe gets
     # "connection refused" -> perpetual unhealthy -> health_monitor restart loop.
     endpoint: http://127.0.0.1:8080
-    path: /
+    # Probe the backend through nginx: a static page can be healthy while
+    # every API/WebSocket request is stuck on a dead backend address.
+    path: /api/v1/backend-info
     interval: 30s
     timeout: 5s
     retries: 3

apps/mempool/manifest.yml

@@ -1,7 +1,7 @@
 app:
   id: mempool
   name: Mempool Explorer
-  version: 3.0.0
+  version: 3.3.1-archy1
   # Where this app comes from, so scripts/check-upstream-releases.py can
   # tell us when the pin below has fallen behind. Without it nothing can:
   # container.image names our mirror, not the project it was mirrored from.
@@ -11,7 +11,7 @@ app:
   description: Bitcoin mempool and blockchain explorer. Real-time transaction and block visualization.
   
   container:
-    image: source.archipelago-foundation.org/lfg2025/mempool-frontend:v3.3.1
+    image: source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1
     image_signature: cosign://...
     pull_policy: if-not-present
     

catalog.json

@@ -67,13 +67,13 @@
     {
       "id": "mempool",
       "title": "Mempool Explorer",
-      "version": "3.0.0",
+      "version": "3.3.1-archy1",
       "description": "Bitcoin mempool and blockchain explorer. Real-time transaction and block visualization.",
       "icon": "/assets/img/app-icons/mempool.webp",
       "author": "Mempool",
       "category": "money",
       "tier": "core",
-      "dockerImage": "source.archipelago-foundation.org/lfg2025/mempool-frontend:v3.3.1",
+      "dockerImage": "source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1",
       "repoUrl": "https://github.com/mempool/mempool",
       "requires": [
         "bitcoin-knots",

core/Cargo.lock

@@ -104,7 +104,7 @@ dependencies = [
 
 [[package]]
 name = "archipelago"
-version = "1.8.14-alpha"
+version = "1.8.17-alpha"
 dependencies = [
  "anyhow",
  "archipelago-container",

core/archipelago/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "archipelago"
-version = "1.8.14-alpha"
+version = "1.8.17-alpha"
 edition = "2021"
 license.workspace = true
 description = "Archipelago Bitcoin Node OS - Native backend"

core/archipelago/src/api/handler/content.rs

@@ -162,11 +162,25 @@ impl ApiHandler {
                     r#"{"error":"This file is shared with the host's federation peers only. Federate with that node (exchange invites) so it recognizes you, then try again."}"#,
                 ),
             )),
-            Ok(content_server::ServeResult::NotFound) | Err(_) => Ok(build_response(
+            Ok(content_server::ServeResult::NotFound) => Ok(build_response(
                 StatusCode::NOT_FOUND,
                 "text/plain",
                 hyper::Body::from("Content not found"),
             )),
+            // A server-side failure is NOT "not found": reporting it as a 404
+            // hid an unreadable file behind a silent, unlogged response, and a
+            // buyer's client re-sends a 404 over another transport. 5xx it, and
+            // say why in the journal.
+            Err(e) => {
+                tracing::warn!(content_id = %content_id, "content request failed: {e:#}");
+                Ok(build_response(
+                    StatusCode::INTERNAL_SERVER_ERROR,
+                    "application/json",
+                    hyper::Body::from(
+                        r#"{"error":"The seller could not read this file right now. You have not been charged."}"#,
+                    ),
+                ))
+            }
         }
     }
 

core/archipelago/src/api/rpc/content.rs

@@ -22,9 +22,11 @@ const FILE_CATALOG_PROTOCOL: &str = "https://archipelago.dev/protocols/file-cata
 /// Best-effort reclaim of an ecash payment token that was minted but the sale
 /// didn't complete (seller unreachable or couldn't redeem it), so the buyer
 /// doesn't lose the value. For Fedimint the spender can reissue its own
-/// un-redeemed notes; for Cashu the proofs are received back. Fails silently if
-/// the seller already claimed the token (then the value is genuinely gone).
-async fn reclaim_spent_ecash(data_dir: &std::path::Path, token: &str, backend: &str) {
+/// un-redeemed notes; for Cashu the proofs are received back. Returns whether
+/// the value came back: false if the seller already claimed the token (then
+/// the value is genuinely gone), so callers never tell the buyer they were
+/// refunded when they weren't.
+async fn reclaim_spent_ecash(data_dir: &std::path::Path, token: &str, backend: &str) -> bool {
     let res = match backend {
         "fedimint" => crate::wallet::fedimint_client::reissue_into_any(data_dir, token)
             .await
@@ -32,13 +34,29 @@ async fn reclaim_spent_ecash(data_dir: &std::path::Path, token: &str, backend: &
         _ => ecash::receive_token(data_dir, token).await,
     };
     match res {
-        Ok(sats) => tracing::info!(
-            "paid download: reclaimed {sats} sats of unspent {backend} ecash after a failed sale"
-        ),
-        Err(e) => tracing::warn!(
-            "paid download: could not reclaim {backend} ecash (the peer may have already \
-             claimed it): {e:#}"
-        ),
+        Ok(sats) => {
+            tracing::info!(
+                "paid download: reclaimed {sats} sats of unspent {backend} ecash after a failed sale"
+            );
+            true
+        }
+        Err(e) => {
+            tracing::warn!(
+                "paid download: could not reclaim {backend} ecash (the peer may have already \
+                 claimed it): {e:#}"
+            );
+            false
+        }
+    }
+}
+
+/// What to tell the buyer about their payment after a failed sale.
+fn refund_note(reclaimed: bool) -> &'static str {
+    if reclaimed {
+        "Your ecash was refunded to your wallet."
+    } else {
+        "The seller had already claimed the payment, so it could not be refunded \
+         automatically — contact the seller."
     }
 }
 
@@ -564,9 +582,13 @@ impl RpcHandler {
                 tracing::warn!("paid peer download dial failed for {}: {:#}", onion, e);
                 // The token was already minted/spent — reclaim it so the buyer
                 // doesn't lose the value when the seller was simply unreachable.
-                reclaim_spent_ecash(&self.config.data_dir, &token_str, used_backend).await;
+                let reclaimed =
+                    reclaim_spent_ecash(&self.config.data_dir, &token_str, used_backend).await;
                 return Ok(serde_json::json!({
-                    "error": "Could not reach the peer over mesh or Tor — it may be offline. Your ecash was refunded to your wallet. Please try again."
+                    "error": format!(
+                        "Could not reach the peer over mesh or Tor — it may be offline. {} Please try again.",
+                        refund_note(reclaimed)
+                    )
                 }));
             }
         };
@@ -592,15 +614,19 @@ impl RpcHandler {
             );
             // Seller couldn't redeem the token — reclaim it so the buyer keeps
             // their funds (the spent-but-unredeemed-notes case the user hit).
-            reclaim_spent_ecash(&self.config.data_dir, &token_str, used_backend).await;
+            let reclaimed =
+                reclaim_spent_ecash(&self.config.data_dir, &token_str, used_backend).await;
+            // The 402 body is generic, so don't assert a cause — a seller that
+            // redeemed the token and then failed to deliver also lands here.
             let hint = match used_backend {
-                "fedimint" => "the seller isn't in the same Fedimint federation as you",
-                _ => "the seller doesn't accept your Cashu mint",
+                "fedimint" => "the seller may not be in the same Fedimint federation as you",
+                _ => "the seller may not accept your Cashu mint",
             };
             return Ok(serde_json::json!({
                 "error": format!(
-                    "Payment rejected by the seller — {hint}. Your ecash was refunded to \
-                     your wallet. Try the other ecash type, or use a shared mint/federation."
+                    "Payment not accepted by the seller — {hint}. {} Try the other ecash \
+                     type, or use a shared mint/federation.",
+                    refund_note(reclaimed)
                 )
             }));
         }
@@ -609,9 +635,10 @@ impl RpcHandler {
             let status = response.status();
             let body = response.text().await.unwrap_or_default();
             tracing::warn!("paid download: seller {onion} returned {status}: {body}");
-            reclaim_spent_ecash(&self.config.data_dir, &token_str, used_backend).await;
+            let reclaimed =
+                reclaim_spent_ecash(&self.config.data_dir, &token_str, used_backend).await;
             return Ok(serde_json::json!({
-                "error": format!("Peer returned an error ({status}). Your ecash was refunded to your wallet.")
+                "error": format!("Peer returned an error ({status}). {}", refund_note(reclaimed))
             }));
         }
 

core/archipelago/src/api/rpc/lnd/info.rs

@@ -73,6 +73,42 @@ struct LndChannelBalanceResponse {
     pending_open_local_balance: Option<LndAmount>,
 }
 
+/// Reject unavailable LND data before it can be decoded as an empty, zero wallet.
+async fn get_lnd_json<T: serde::de::DeserializeOwned>(
+    client: &reqwest::Client,
+    url: &str,
+    macaroon_hex: &str,
+) -> Result<T> {
+    client
+        .get(url)
+        .header("Grpc-Metadata-macaroon", macaroon_hex)
+        .send()
+        .await
+        .context("LND is unavailable; balance could not be checked")?
+        .error_for_status()
+        .context("LND is not ready; balance could not be checked")?
+        .json()
+        .await
+        .context("LND returned invalid wallet data")
+}
+
+fn checked_balances(
+    wallet: LndBalanceResponse,
+    channels: LndChannelBalanceResponse,
+) -> Result<(i64, i64, i64)> {
+    fn sats(value: Option<String>) -> Result<i64> {
+        let value = value.context("LND omitted a balance; balance is unavailable")?;
+        let amount: i64 = value.parse().context("LND returned an invalid balance")?;
+        anyhow::ensure!(amount >= 0, "LND returned a negative balance");
+        Ok(amount)
+    }
+    Ok((
+        sats(wallet.total_balance)?,
+        sats(channels.local_balance.and_then(|a| a.sat))?,
+        sats(channels.pending_open_local_balance.and_then(|a| a.sat))?,
+    ))
+}
+
 impl RpcHandler {
     pub(in crate::api::rpc) async fn handle_lnd_getinfo(&self) -> Result<serde_json::Value> {
         let macaroon_bytes = read_lnd_admin_macaroon().await?;
@@ -85,45 +121,26 @@ impl RpcHandler {
             .build()
             .context("Failed to create HTTP client")?;
 
-        let get_info: LndGetInfoResponse = client
-            .get(format!("{LND_REST_BASE_URL}/v1/getinfo"))
-            .header("Grpc-Metadata-macaroon", &macaroon_hex)
-            .send()
-            .await
-            .context("LND REST connection failed")?
-            .json()
-            .await
-            .context("Failed to parse LND getinfo response")?;
-
-        let channel_balance: LndChannelBalanceResponse = match client
-            .get(format!("{LND_REST_BASE_URL}/v1/balance/channels"))
-            .header("Grpc-Metadata-macaroon", &macaroon_hex)
-            .send()
-            .await
-        {
-            Ok(resp) => resp.json().await.unwrap_or(LndChannelBalanceResponse {
-                local_balance: None,
-                pending_open_local_balance: None,
-            }),
-            Err(_) => LndChannelBalanceResponse {
-                local_balance: None,
-                pending_open_local_balance: None,
-            },
-        };
-
-        let wallet_balance: LndBalanceResponse = match client
-            .get(format!("{LND_REST_BASE_URL}/v1/balance/blockchain"))
-            .header("Grpc-Metadata-macaroon", &macaroon_hex)
-            .send()
-            .await
-        {
-            Ok(resp) => resp.json().await.unwrap_or(LndBalanceResponse {
-                total_balance: None,
-            }),
-            Err(_) => LndBalanceResponse {
-                total_balance: None,
-            },
-        };
+        let get_info: LndGetInfoResponse = get_lnd_json(
+            &client,
+            &format!("{LND_REST_BASE_URL}/v1/getinfo"),
+            &macaroon_hex,
+        )
+        .await?;
+        let channel_balance: LndChannelBalanceResponse = get_lnd_json(
+            &client,
+            &format!("{LND_REST_BASE_URL}/v1/balance/channels"),
+            &macaroon_hex,
+        )
+        .await?;
+        let wallet_balance: LndBalanceResponse = get_lnd_json(
+            &client,
+            &format!("{LND_REST_BASE_URL}/v1/balance/blockchain"),
+            &macaroon_hex,
+        )
+        .await?;
+        let (balance_sats, channel_balance_sats, pending_open_balance) =
+            checked_balances(wallet_balance, channel_balance)?;
 
         let (identity_pubkey, uris) = map_identity(&get_info);
 
@@ -135,18 +152,9 @@ impl RpcHandler {
             num_peers: get_info.num_peers.unwrap_or(0),
             synced_to_chain: get_info.synced_to_chain.unwrap_or(false),
             block_height: get_info.block_height.unwrap_or(0),
-            balance_sats: wallet_balance
-                .total_balance
-                .and_then(|s| s.parse().ok())
-                .unwrap_or(0),
-            channel_balance_sats: channel_balance
-                .local_balance
-                .and_then(|a| a.sat.and_then(|s| s.parse().ok()))
-                .unwrap_or(0),
-            pending_open_balance: channel_balance
-                .pending_open_local_balance
-                .and_then(|a| a.sat.and_then(|s| s.parse().ok()))
-                .unwrap_or(0),
+            balance_sats,
+            channel_balance_sats,
+            pending_open_balance,
         };
 
         Ok(serde_json::to_value(info)?)
@@ -268,6 +276,76 @@ impl RpcHandler {
 mod tests {
     use super::*;
 
+    #[test]
+    fn unavailable_balances_are_not_zero() {
+        for body in [r#"{}"#, r#"{"code":14,"message":"wallet locked"}"#] {
+            assert!(checked_balances(
+                serde_json::from_str(body).unwrap(),
+                serde_json::from_str(body).unwrap(),
+            )
+            .is_err());
+        }
+        for value in ["bad", "-1", "9223372036854775808"] {
+            let wallet = LndBalanceResponse {
+                total_balance: Some(value.into()),
+            };
+            let channels = serde_json::from_str(
+                r#"{"local_balance":{"sat":"5"},"pending_open_local_balance":{"sat":"0"}}"#,
+            )
+            .unwrap();
+            assert!(checked_balances(wallet, channels).is_err());
+        }
+    }
+
+    #[test]
+    fn verified_zero_and_nonzero_balances_survive() {
+        for expected in [0, 42] {
+            let wallet = LndBalanceResponse {
+                total_balance: Some(expected.to_string()),
+            };
+            let channels = serde_json::from_value(serde_json::json!({
+                "local_balance":{"sat":expected.to_string()},
+                "pending_open_local_balance":{"sat":"0"}
+            }))
+            .unwrap();
+            assert_eq!(
+                checked_balances(wallet, channels).unwrap(),
+                (expected, expected, 0)
+            );
+        }
+    }
+
+    #[tokio::test]
+    async fn locked_wallet_http_response_is_not_successful_getinfo() {
+        use tokio::io::{AsyncReadExt, AsyncWriteExt};
+        let listener = tokio::net::TcpListener::bind("127.0.0.1:0").await.unwrap();
+        let addr = listener.local_addr().unwrap();
+        let server = tokio::spawn(async move {
+            let (mut stream, _) = listener.accept().await.unwrap();
+            let mut buf = [0; 2048];
+            stream.read(&mut buf).await.unwrap();
+            let body =
+                r#"{"code":9,"message":"wallet locked, unlock it to enable full RPC access"}"#;
+            stream.write_all(format!(
+                "HTTP/1.1 503 Service Unavailable\r\nContent-Type: application/json\r\nContent-Length: {}\r\nConnection: close\r\n\r\n{}",
+                body.len(), body
+            ).as_bytes()).await.unwrap();
+        });
+        let client = reqwest::Client::builder()
+            .no_proxy()
+            .timeout(std::time::Duration::from_secs(2))
+            .build()
+            .unwrap();
+        assert!(get_lnd_json::<LndGetInfoResponse>(
+            &client,
+            &format!("http://{addr}/v1/getinfo"),
+            "test"
+        )
+        .await
+        .is_err());
+        server.await.unwrap();
+    }
+
     /// A real compressed secp256k1 pubkey shape: 66 hex characters.
     const GOOD_PUBKEY: &str = "03a1b2c3d4e5f60718293a4b5c6d7e8f90a1b2c3d4e5f60718293a4b5c6d7e8f90";
 

core/archipelago/src/api/rpc/package/async_lifecycle.rs

@@ -330,7 +330,7 @@ impl RpcHandler {
         let package_id_spawn = package_id.clone();
         tokio::spawn(async move {
             match handler.handle_package_update(params).await {
-                Ok(_) => {
+                Ok(result) => {
                     info!("package.update {}: complete", package_id_spawn);
                     // Same reasoning as install: the merge_preserving_transitional
                     // helper treats Updating as RPC-owned, so we MUST write the
@@ -345,7 +345,11 @@ impl RpcHandler {
                     set_package_state(
                         &handler.state_manager,
                         &package_id_spawn,
-                        PackageState::Running,
+                        if result.get("status").and_then(|v| v.as_str()) == Some("up-to-date") {
+                            pre_state.clone().unwrap_or(PackageState::Running)
+                        } else {
+                            PackageState::Running
+                        },
                     )
                     .await;
                 }

core/archipelago/src/api/rpc/package/update.rs

@@ -19,7 +19,7 @@ use tracing::{error, info, warn};
 const PODMAN_UPDATE_PULL_TIMEOUT: std::time::Duration = std::time::Duration::from_secs(600);
 
 impl RpcHandler {
-    /// Update a package to the version pinned in image-versions.sh.
+    /// Update a package to the freshly verified catalog target.
     /// This is a manual operation — the user clicks "Update" in the UI.
     pub(in crate::api::rpc) async fn handle_package_update(
         &self,
@@ -32,6 +32,21 @@ impl RpcHandler {
             .ok_or_else(|| anyhow::anyhow!("Missing package id"))?;
         validate_app_id(package_id)?;
 
+        // An Update click must not act on an hourly cache that predates the
+        // button. Fetch and verify first; failure leaves running containers alone.
+        crate::container::app_catalog::refresh_catalog(&self.config.data_dir)
+            .await
+            .context(
+                "Cannot check the signed app catalog; update cancelled before changing containers",
+            )?;
+        if let Some(orch) = &self.orchestrator {
+            // Reload even when bytes did not change: a previous reload may have
+            // failed after the cache was written, or another refresher wrote it.
+            orch.reload_manifests()
+                .await
+                .context("Cannot load current app manifests; update cancelled")?;
+        }
+
         // Resolve the target image. Prefer the remote app catalog (decoupled
         // from the binary OTA), falling back to the image-versions.sh pin. This
         // is OPTIONAL for orchestrator-managed apps: the orchestrator resolves
@@ -42,6 +57,22 @@ impl RpcHandler {
         let pinned = crate::container::app_catalog::catalog_primary_image(package_id)
             .or_else(|| image_versions::pinned_image_for_app(package_id));
 
+        let targets = pinned
+            .as_ref()
+            .map(|target| self.resolve_images_to_pull(package_id, target));
+        if let Some(targets) = &targets {
+            let installed = inspect_update_images(package_id).await?;
+            if !update_targets_need_change(targets, &installed)? {
+                install_log(&format!(
+                    "UPDATE SKIP: {} — target versions already installed",
+                    package_id
+                ))
+                .await;
+                self.clear_install_progress(package_id).await;
+                return Ok(serde_json::json!({"status": "up-to-date", "package_id": package_id}));
+            }
+        }
+
         // Note: the `already updating` guard lives in `spawn_package_update`
         // (the async wrapper that dispatch actually routes to). By the time
         // this inner function runs, the wrapper has already flipped state to
@@ -80,6 +111,12 @@ impl RpcHandler {
             if let Some(orchestrator) = self.orchestrator.as_ref() {
                 match orchestrator.upgrade(orchestrator_app_id).await {
                     Ok(()) => {
+                        if let Some(targets) = &targets {
+                            verify_update_targets(
+                                targets,
+                                &inspect_update_images(package_id).await?,
+                            )?;
+                        }
                         self.set_install_phase(package_id, InstallPhase::WaitingHealthy)
                             .await;
                         if let Ok(health) = orchestrator.health(orchestrator_app_id).await {
@@ -133,7 +170,8 @@ impl RpcHandler {
         };
 
         // Resolve images to pull — either a stack or single container
-        let images_to_pull = self.resolve_images_to_pull(package_id, &pinned);
+        let images_to_pull =
+            targets.unwrap_or_else(|| self.resolve_images_to_pull(package_id, &pinned));
 
         // Get all containers for this app
         let containers = get_containers_for_app(package_id).await?;
@@ -324,15 +362,22 @@ impl RpcHandler {
                 .await;
             if let Ok(o) = status {
                 let state = String::from_utf8_lossy(&o.stdout).trim().to_string();
-                if state == "exited" {
-                    warn!(
-                        "Update {}: container {} exited after recreate",
-                        package_id, name
-                    );
-                }
+                anyhow::ensure!(
+                    o.status.success() && state == "running",
+                    "Update {}: container {} is not running after recreate",
+                    package_id,
+                    name
+                );
+            } else {
+                anyhow::bail!(
+                    "Update {}: cannot inspect recreated container {}",
+                    package_id,
+                    name
+                );
             }
         }
 
+        verify_update_targets(images_to_pull, &inspect_update_images(package_id).await?)?;
         Ok(())
     }
 
@@ -514,6 +559,98 @@ impl RpcHandler {
     }
 }
 
+async fn inspect_update_images(package_id: &str) -> Result<Vec<(String, String)>> {
+    let containers = get_containers_for_app(package_id).await?;
+    anyhow::ensure!(
+        !containers.is_empty(),
+        "No containers found for {}",
+        package_id
+    );
+    let mut command = tokio::process::Command::new("podman");
+    command.arg("inspect").args(&containers).kill_on_drop(true);
+    let output = tokio::time::timeout(std::time::Duration::from_secs(30), command.output())
+        .await
+        .context("Timed out checking installed images")??;
+    anyhow::ensure!(
+        output.status.success(),
+        "Cannot inspect installed images; update cancelled"
+    );
+    let inspected: Vec<serde_json::Value> = serde_json::from_slice(&output.stdout)?;
+    inspected
+        .iter()
+        .map(|entry| {
+            let name = entry
+                .get("Name")
+                .and_then(|v| v.as_str())
+                .ok_or_else(|| anyhow::anyhow!("Container inspection omitted Name"))?;
+            let image = entry
+                .get("ImageName")
+                .and_then(|v| v.as_str())
+                .ok_or_else(|| anyhow::anyhow!("Container inspection omitted ImageName"))?;
+            Ok((name.trim_start_matches('/').to_string(), image.to_string()))
+        })
+        .collect()
+}
+
+fn installed_image_for_target<'a>(
+    app_id: &str,
+    installed: &'a [(String, String)],
+) -> Option<&'a str> {
+    installed
+        .iter()
+        .find(|(name, _)| {
+            candidate_app_ids_for_container(name)
+                .iter()
+                .any(|id| id == app_id)
+        })
+        .map(|(_, image)| image.as_str())
+}
+
+/// A successful recreate is not proof that it used the downloaded image.
+fn verify_update_targets(
+    targets: &[(String, String)],
+    installed: &[(String, String)],
+) -> Result<()> {
+    for (app_id, target) in targets {
+        let running = installed_image_for_target(app_id, installed).ok_or_else(|| {
+            anyhow::anyhow!("Update {}: target container missing after recreate", app_id)
+        })?;
+        anyhow::ensure!(
+            image_versions::extract_version_from_image(target)
+                == image_versions::extract_version_from_image(running)
+                || image_versions::compare_image_versions(target, running)
+                    == Some(std::cmp::Ordering::Equal),
+            "Update {}: recreated container did not reach target version {}",
+            app_id,
+            image_versions::extract_version_from_image(target)
+        );
+    }
+    Ok(())
+}
+
+/// Check every stack component, not just the version shown on its tile. A
+/// newer backend must still update when its frontend version is unchanged.
+/// A stale target for any component cancels before pulling or stopping anything.
+fn update_targets_need_change(
+    targets: &[(String, String)],
+    installed: &[(String, String)],
+) -> Result<bool> {
+    use std::cmp::Ordering;
+    let mut changed = false;
+    for (app_id, target) in targets {
+        let running = installed_image_for_target(app_id, installed);
+        match running.and_then(|image| image_versions::compare_image_versions(target, image)) {
+            Some(Ordering::Less) => anyhow::bail!(
+                "Catalog target for {} is older than the installed image; refusing downgrade",
+                app_id
+            ),
+            Some(Ordering::Equal) => {}
+            Some(Ordering::Greater) | None => changed = true,
+        }
+    }
+    Ok(changed)
+}
+
 fn should_try_orchestrator_update(package_id: &str, orchestrator_available: bool) -> bool {
     orchestrator_available && !uses_legacy_update_flow(package_id)
 }
@@ -526,11 +663,14 @@ fn orchestrator_update_app_id(package_id: &str) -> &str {
 }
 
 fn uses_legacy_update_flow(package_id: &str) -> bool {
-    matches!(
-        package_id,
-        // Multi-container stacks still updated via the stack-aware path.
-        "immich" | "penpot" | "penpot-frontend" | "indeedhub"
-    )
+    // A primary container already at its target does not mean its backend or
+    // database is current. Route every mapped stack through the component flow.
+    !image_versions::containers_for_stack(package_id).is_empty()
+        || matches!(
+            package_id,
+            // Multi-container stacks still updated via the stack-aware path.
+            "immich" | "penpot" | "penpot-frontend" | "indeedhub"
+        )
 }
 
 fn is_unknown_app_id_error(err: &anyhow::Error) -> bool {
@@ -554,7 +694,12 @@ fn candidate_app_ids_for_container(container_name: &str) -> Vec<String> {
         "archy-bitcoin-ui" => push("bitcoin-ui"),
         "archy-lnd-ui" => push("lnd-ui"),
         "archy-electrs-ui" => push("electrs-ui"),
-        "mempool" => {
+        "mysql-mempool" => push("archy-mempool-db"),
+        "btcpay" | "btcpayserver" | "archy-btcpay" => push("btcpay-server"),
+        "homeassistant" | "archy-homeassistant" => push("home-assistant"),
+        "fedimintd" => push("fedimint"),
+        "electrs" | "mempool-electrs" => push("electrumx"),
+        "mempool" | "mempool-web" => {
             push("archy-mempool-web");
             push("mempool");
         }
@@ -572,27 +717,89 @@ fn candidate_app_ids_for_container(container_name: &str) -> Vec<String> {
 mod tests {
     use super::{
         candidate_app_ids_for_container, orchestrator_update_app_id,
-        should_try_orchestrator_update, uses_legacy_update_flow,
+        should_try_orchestrator_update, update_targets_need_change, uses_legacy_update_flow,
+        verify_update_targets,
     };
 
+    #[test]
+    fn mempool_update_preflight_rejects_stale_catalog_without_reinstalling() {
+        let installed = vec![(
+            "mempool".into(),
+            "r.test/lfg2025/mempool-frontend:v3.3.1-archy1".into(),
+        )];
+        let stale = vec![(
+            "archy-mempool-web".into(),
+            "r.test/lfg2025/mempool-frontend:v3.3.1".into(),
+        )];
+        assert!(update_targets_need_change(&stale, &installed).is_err());
+        let current = vec![(
+            "archy-mempool-web".into(),
+            "r.test/chaum/mempool-frontend:v3.3.1-archy1".into(),
+        )];
+        assert!(!update_targets_need_change(&current, &installed).unwrap());
+        let legacy = vec![(
+            "mempool-web".into(),
+            "r.test/old/mempool-frontend:v3.3.1-archy1".into(),
+        )];
+        assert!(!update_targets_need_change(&current, &legacy).unwrap());
+        let newer = vec![(
+            "archy-mempool-web".into(),
+            "r.test/chaum/mempool-frontend:v3.3.1-archy2".into(),
+        )];
+        assert!(update_targets_need_change(&newer, &installed).unwrap());
+    }
+
+    #[test]
+    fn stack_update_checks_backend_even_when_frontend_matches() {
+        let installed = vec![
+            ("mempool".into(), "r.test/team/web:3.3.1-archy1".into()),
+            ("mempool-api".into(), "r.test/team/api:3.3.1".into()),
+        ];
+        let mut targets = vec![
+            (
+                "archy-mempool-web".into(),
+                "r.test/team/web:3.3.1-archy1".into(),
+            ),
+            ("mempool-api".into(), "r.test/team/api:3.3.2".into()),
+        ];
+        assert!(update_targets_need_change(&targets, &installed).unwrap());
+        targets[0].1 = "r.test/team/web:3.3.1".into();
+        assert!(update_targets_need_change(&targets, &installed).is_err());
+    }
+
+    #[test]
+    fn update_completion_requires_the_target_version_to_be_installed() {
+        let targets = vec![(
+            "archy-mempool-web".into(),
+            "r.test/chaum/mempool-frontend:v3.3.1-archy1".into(),
+        )];
+        let mut installed = vec![(
+            "mempool".into(),
+            "r.test/lfg2025/mempool-frontend:v3.3.1".into(),
+        )];
+        assert!(verify_update_targets(&targets, &installed).is_err());
+        assert!(verify_update_targets(&targets, &[]).is_err());
+        installed[0].1 = "r.test/lfg2025/mempool-frontend:v3.3.1-archy1".into();
+        assert!(verify_update_targets(&targets, &installed).is_ok());
+    }
+
     #[test]
     fn legacy_flow_for_stack_apps() {
-        for app in ["immich", "penpot", "indeedhub"] {
+        for app in [
+            "immich",
+            "penpot",
+            "indeedhub",
+            "mempool",
+            "btcpay-server",
+            "netbird",
+        ] {
             assert!(uses_legacy_update_flow(app), "{app} should stay legacy");
         }
     }
 
     #[test]
     fn orchestrator_flow_for_single_apps() {
-        for app in [
-            "lnd",
-            "bitcoin-core",
-            "searxng",
-            "grafana",
-            "btcpay-server",
-            "mempool",
-            "fedimint",
-        ] {
+        for app in ["lnd", "bitcoin-core", "searxng", "grafana", "fedimint"] {
             assert!(
                 !uses_legacy_update_flow(app),
                 "{app} should be orchestrator-first"

core/archipelago/src/container/app_catalog.rs

@@ -400,7 +400,7 @@ pub fn available_update_for_app(app_id: &str, running_image: &str) -> Option<Str
     }
     if let Some(catalog_image) = catalog_primary_image(app_id) {
         // Catalog covers this app with a concrete image -> authoritative.
-        return crate::container::image_versions::available_update_for_images(
+        return crate::container::image_versions::available_catalog_update_for_images(
             &catalog_image,
             running_image,
         );

core/archipelago/src/container/image_versions.rs

@@ -100,6 +100,12 @@ fn parse_image_versions(content: &str) -> HashMap<String, String> {
 
         // Match VAR="value" or VAR=value
         if let Some((key, val)) = parse_assignment(line) {
+            // Read a self-default assignment without evaluating shell code.
+            let default_prefix = format!("${{{key}:-");
+            let val = val
+                .strip_prefix(&default_prefix)
+                .and_then(|v| v.strip_suffix('}'))
+                .unwrap_or(val);
             let expanded = val.replace("$ARCHY_REGISTRY", &registry);
             if key == "ARCHY_REGISTRY" {
                 registry = expanded.clone();
@@ -205,48 +211,71 @@ pub fn available_update_for_app(app_id: &str, running_image: &str) -> Option<Str
 }
 
 pub fn available_update_for_images(pinned: &str, running_image: &str) -> Option<String> {
-    let pinned_version = extract_version_from_image(&pinned);
+    if image_without_registry_or_tag(pinned) != image_without_registry_or_tag(running_image) {
+        return None;
+    }
+    available_catalog_update_for_images(pinned, running_image)
+}
+
+/// A signed catalog binds the image to an app id, so a publisher namespace
+/// migration must not hide a real upgrade. Baseline pins still require the
+/// same repository via `available_update_for_images` above.
+pub fn available_catalog_update_for_images(pinned: &str, running_image: &str) -> Option<String> {
+    let pinned_version = extract_version_from_image(pinned);
     if is_floating_tag(&pinned_version) {
         return None;
     }
-
-    let running_version = extract_version_from_image(running_image);
-    if pinned_version == running_version {
-        return None;
-    }
-
-    let pinned_repo = image_without_registry_or_tag(&pinned);
-    let running_repo = image_without_registry_or_tag(running_image);
-    if pinned_repo != running_repo {
-        return None;
-    }
-
-    // Never advertise a LOWER version as an update.
-    //
-    // Everything upstream of here is a version claim that can go stale: the
-    // signed catalog, a legacy catalog entry with no manifest, the
-    // image-versions.sh baseline pin. When one lags behind what a node is
-    // actually running, a bare `pinned != running` check turns that staleness
-    // into an "Update" button that rolls the node BACKWARDS — and a rollback
-    // to a version withdrawn for a vulnerability is precisely the case where
-    // that must not happen. Observed with BTCPay: 2.4.2 installed, a stale
-    // 2.3.9 pin, and the UI offering "update" to the exploited release.
-    //
-    // Only suppress when both tags parse as comparable version numbers, so
-    // apps with opaque tags (RELEASE.2024-11-07T00-52-20Z, 14-vectorchord0.4.3)
-    // keep the previous behaviour rather than silently losing updates.
-    if let (Some(p), Some(r)) = (
-        parse_version_parts(&pinned_version),
-        parse_version_parts(&running_version),
+    if matches!(
+        compare_image_versions(pinned, running_image),
+        Some(std::cmp::Ordering::Less | std::cmp::Ordering::Equal)
     ) {
-        if p < r {
-            return None;
-        }
+        return None;
     }
-
     Some(pinned_version)
 }
 
+/// Compare explicit image tags, ignoring registry and namespace. `None` means
+/// unknown ordering (including floating tags), never permission to downgrade.
+/// Archipelago's `-archyN` is a downstream patch revision ABOVE the upstream
+/// release, not a SemVer prerelease below it.
+pub fn compare_image_versions(target: &str, running: &str) -> Option<std::cmp::Ordering> {
+    use std::cmp::Ordering;
+    let target = extract_version_from_image(target);
+    let running = extract_version_from_image(running);
+    if is_floating_tag(&target) || is_floating_tag(&running) {
+        return None;
+    }
+    let target = target.strip_prefix('v').unwrap_or(&target);
+    let running = running.strip_prefix('v').unwrap_or(&running);
+    if target == running {
+        return Some(Ordering::Equal);
+    }
+    let mut target_core = parse_version_parts(target)?;
+    let mut running_core = parse_version_parts(running)?;
+    while target_core.last() == Some(&0) {
+        target_core.pop();
+    }
+    while running_core.last() == Some(&0) {
+        running_core.pop();
+    }
+    match target_core.cmp(&running_core) {
+        Ordering::Equal => {
+            fn patch_revision(tag: &str) -> Option<u64> {
+                if let Some((base, revision)) = tag.rsplit_once("-archy") {
+                    if base.chars().all(|c| c.is_ascii_digit() || c == '.') {
+                        return revision.parse().ok();
+                    }
+                }
+                tag.chars()
+                    .all(|c| c.is_ascii_digit() || c == '.')
+                    .then_some(0)
+            }
+            Some(patch_revision(target)?.cmp(&patch_revision(running)?))
+        }
+        order => Some(order),
+    }
+}
+
 /// Numeric components of a version tag, for ordering comparisons only.
 ///
 /// Accepts a leading `v` and a trailing pre-release suffix (`v0.18.4-beta`),
@@ -423,6 +452,57 @@ mod tests {
         );
     }
 
+    #[test]
+    fn downstream_patch_is_newer_than_upstream_and_orders_revisions() {
+        let upstream = "registry.test/team/mempool-frontend:v3.3.1";
+        let patch1 = "registry.test/team/mempool-frontend:v3.3.1-archy1";
+        let patch2 = "registry.test/team/mempool-frontend:v3.3.1-archy2";
+        assert_eq!(available_update_for_images(upstream, patch1), None);
+        assert_eq!(available_update_for_images(patch1, patch2), None);
+        assert_eq!(
+            available_update_for_images(patch1, upstream),
+            Some("v3.3.1-archy1".into())
+        );
+        assert_eq!(
+            available_update_for_images(patch2, patch1),
+            Some("v3.3.1-archy2".into())
+        );
+    }
+
+    #[test]
+    fn catalog_namespace_migration_does_not_hide_patch_or_offer_reinstall() {
+        let old = "registry.test/lfg2025/mempool-frontend:v3.3.1";
+        let patched = "registry.test/chaum/mempool-frontend:v3.3.1-archy1";
+        assert_eq!(
+            available_catalog_update_for_images(patched, old),
+            Some("v3.3.1-archy1".into())
+        );
+        assert_eq!(
+            available_catalog_update_for_images(
+                patched,
+                "registry.test/lfg2025/mempool-frontend:v3.3.1-archy1"
+            ),
+            None
+        );
+        assert_eq!(available_update_for_images(patched, old), None);
+    }
+
+    #[test]
+    fn equivalent_version_spelling_does_not_offer_update() {
+        assert_eq!(
+            available_update_for_images("r.test/team/app:v3.3.1", "r.test/team/app:3.3.1"),
+            None
+        );
+        assert_eq!(
+            available_update_for_images("r.test/team/app:3.3.0", "r.test/team/app:3.3"),
+            None
+        );
+        assert_eq!(
+            compare_image_versions("r.test/team/app:latest", "r.test/team/app:latest"),
+            None
+        );
+    }
+
     #[test]
     fn test_parse_image_versions() {
         let content = r#"
@@ -445,6 +525,22 @@ NOT_AN_IMAGE="something"
         assert!(!parsed.contains_key("ARCHY_REGISTRY"));
     }
 
+    #[test]
+    fn shipped_image_pins_expand_shell_defaults_to_concrete_refs() {
+        let images = parse_image_versions(include_str!("../../../../scripts/image-versions.sh"));
+        assert_eq!(
+            images["MEMPOOL_WEB_IMAGE"],
+            "source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1"
+        );
+        assert_eq!(
+            images["MEMPOOL_BACKEND_IMAGE"],
+            "source.archipelago-foundation.org/lfg2025/mempool-backend:v3.3.1"
+        );
+        assert!(images
+            .values()
+            .all(|v| !v.contains('$') && !v.contains('}')));
+    }
+
     #[test]
     fn test_image_var_mapping() {
         assert_eq!(image_var_for_app("lnd"), Some("LND_IMAGE"));

core/archipelago/src/container/lnd.rs

@@ -96,129 +96,21 @@ pub async fn ensure_wallet_initialized() -> Result<()> {
         if file_exists_as_root(admin_macaroon).await && lnd_getinfo_ready(admin_macaroon).await {
             return Ok(());
         }
-        match unlock_existing_wallet().await? {
-            true => {
-                wait_for_admin_macaroon(admin_macaroon).await?;
-                return Ok(());
-            }
-            false => {
-                // Every candidate password was actively rejected: this wallet was
-                // created with a password this node no longer has, so it can never
-                // auto-unlock unattended. Alpha nodes hold no real funds and a wallet
-                // locked with an unknown password is already inaccessible, so wipe +
-                // recreate it on the per-node secret to self-heal at boot.
-                recreate_wallet_destructively().await?;
-                wait_for_admin_macaroon(admin_macaroon).await?;
-                return Ok(());
-            }
-        }
+        unlock_existing_wallet_no_wipe().await?;
+        wait_for_admin_macaroon(admin_macaroon).await?;
+        return Ok(());
     }
 
     init_wallet_via_rest().await?;
     wait_for_admin_macaroon(admin_macaroon).await
 }
 
-/// LND data subdirectories holding wallet + channel + graph state. Removing them
-/// returns LND to a NON_EXISTING wallet state. Funds-bearing data lives here too,
-/// so deletion is destructive — only done once the wallet is already unrecoverable.
-const LND_STATE_DIRS: &[&str] = &[
-    "/var/lib/archipelago/lnd/data/chain",
-    "/var/lib/archipelago/lnd/data/graph",
-];
-
-/// Podman container name for the core LND app (see `compute_container_name`:
-/// non-UI core apps keep their bare id). LND runs as a plain bridge-network
-/// container, not a Quadlet unit, so it is restarted via `podman`, not systemctl.
-const LND_CONTAINER: &str = "lnd";
-
 /// Canonical on-host admin macaroon — same path the RPC layer reads.
 const LND_ADMIN_MACAROON: &str =
     "/var/lib/archipelago/lnd/data/chain/bitcoin/mainnet/admin.macaroon";
 
-/// Archipelago data dir (default; not overridden in prod). Holds the
-/// `user-stopped.json` that gates health-monitor auto-restart.
 const ARCHY_DATA_DIR: &str = "/var/lib/archipelago";
 
-/// Destroy an unrecoverable LND wallet and recreate a fresh one keyed to the
-/// per-node secret. Suppresses health-monitor auto-restart for the wipe window,
-/// stops LND, deletes its wallet/chain/graph state as root, restarts it, waits
-/// for NON_EXISTING, then inits a fresh wallet. Destructive — only called when no
-/// candidate password can open the existing wallet.
-async fn recreate_wallet_destructively() -> Result<()> {
-    tracing::warn!(
-        "[lnd] wallet is locked with an unknown password and cannot auto-unlock; \
-         wiping and recreating it on the per-node secret (DESTRUCTIVE)"
-    );
-
-    // The health monitor restarts any container it sees stopped; mark LND
-    // user-stopped so it doesn't re-launch (and re-open the wallet) mid-wipe.
-    // Always cleared below so LND auto-recovers normally afterwards.
-    let data_dir = std::path::Path::new(ARCHY_DATA_DIR);
-    crate::crash_recovery::mark_user_stopped(data_dir, LND_CONTAINER).await;
-    let result = wipe_and_reinit_wallet().await;
-    crate::crash_recovery::clear_user_stopped(data_dir, LND_CONTAINER).await;
-    result
-}
-
-async fn wipe_and_reinit_wallet() -> Result<()> {
-    podman_user_scoped(&["stop", LND_CONTAINER])
-        .await
-        .context("stopping lnd before wallet wipe")?;
-
-    for dir in LND_STATE_DIRS {
-        let status = host_sudo(&["rm", "-rf", dir])
-            .await
-            .with_context(|| format!("removing {dir}"))?;
-        if !status.success() {
-            anyhow::bail!("removing {dir} exited with {status}");
-        }
-    }
-
-    podman_user_scoped(&["start", LND_CONTAINER])
-        .await
-        .context("restarting lnd after wallet wipe")?;
-
-    wait_for_wallet_state("NON_EXISTING").await?;
-    init_wallet_via_rest().await
-}
-
-/// Run `podman <args>` inside a transient `systemd-run --user --scope`, matching
-/// how the orchestrator/health-monitor manage rootless containers (keeps the
-/// container out of the archipelago service's cgroup).
-async fn podman_user_scoped(args: &[&str]) -> Result<()> {
-    let out = tokio::process::Command::new("systemd-run")
-        .args(["--user", "--scope", "--quiet", "--collect", "podman"])
-        .args(args)
-        .output()
-        .await
-        .with_context(|| format!("systemd-run --user --scope podman {}", args.join(" ")))?;
-    if !out.status.success() {
-        anyhow::bail!(
-            "podman {} failed: {}",
-            args.join(" "),
-            String::from_utf8_lossy(&out.stderr).trim()
-        );
-    }
-    Ok(())
-}
-
-/// Poll `/v1/state` until LND reports `target`, or time out after ~120s.
-async fn wait_for_wallet_state(target: &str) -> Result<()> {
-    let client = reqwest::Client::builder()
-        .no_proxy()
-        .timeout(std::time::Duration::from_secs(5))
-        .danger_accept_invalid_certs(true)
-        .build()
-        .context("building LND REST client")?;
-    for _ in 0..120 {
-        if wallet_state(&client).await.as_deref() == Some(target) {
-            return Ok(());
-        }
-        tokio::time::sleep(std::time::Duration::from_secs(1)).await;
-    }
-    anyhow::bail!("LND did not reach state {target} after wallet wipe")
-}
-
 async fn file_exists_as_root(path: &str) -> bool {
     if std::path::Path::new(path).exists() {
         return true;
@@ -390,14 +282,8 @@ async fn unlock_existing_wallet_via_rest() -> Result<bool> {
     )
 }
 
-/// Unlock an existing wallet WITHOUT the destructive fallback.
-///
-/// `ensure_wallet_initialized` wipes and recreates a wallet no candidate
-/// password can open — correct for a boot path that must self-heal, and exactly
-/// wrong for macaroon rotation, which restarts LND against a wallet the operator
-/// still wants. Rotation calls this instead, so there is no code path from
-/// "rotate my credentials" to "delete my wallet": a rejected password surfaces
-/// as an error the caller reports, never as a wipe.
+/// Unlock the existing wallet, preserving its identity and channel data when
+/// passwords are unavailable or rejected. Used by boot and credential rotation.
 pub(crate) async fn unlock_existing_wallet_no_wipe() -> Result<()> {
     match unlock_existing_wallet().await? {
         true => Ok(()),
@@ -538,7 +424,7 @@ async fn init_wallet_via_rest() -> Result<()> {
     {
         UnlockerResponse::Value(seed) => seed,
         UnlockerResponse::WalletAlreadyExists => {
-            unlock_existing_wallet().await?;
+            unlock_existing_wallet_no_wipe().await?;
             return Ok(());
         }
     };
@@ -569,7 +455,7 @@ async fn init_wallet_via_rest() -> Result<()> {
             .await;
         }
         UnlockerResponse::WalletAlreadyExists => {
-            unlock_existing_wallet().await?;
+            unlock_existing_wallet_no_wipe().await?;
         }
     }
 

core/archipelago/src/container/prod_orchestrator.rs

@@ -1864,7 +1864,7 @@ impl ProdContainerOrchestrator {
         // Durable installation record, consulted alongside the perishable
         // `was_running` snapshot for desired-state recovery below.
         let installed_apps = crate::crash_recovery::load_installed_apps(&self.data_dir).await;
-        let (manifests, container_name_by_app_id): (
+        let (mut manifests, container_name_by_app_id): (
             Vec<LoadedManifest>,
             std::collections::HashMap<String, String>,
         ) = {
@@ -1895,6 +1895,15 @@ impl ProdContainerOrchestrator {
                 .collect();
             (filtered, names)
         };
+        // Wallet readiness must not wait behind unrelated image pulls/builds.
+        // A running LND container can still be locked after boot; its post-start
+        // hook must run promptly. Reconcile Bitcoin first, then LND, before the
+        // rest of the catalog. Each app still honors stopped/uninstalled markers.
+        manifests.sort_by_key(|lm| match lm.manifest.app.id.as_str() {
+            "bitcoin-knots" | "bitcoin-core" | "bitcoin" => 0,
+            "lnd" => 1,
+            _ => 2,
+        });
         // Live container names (any state), for the same recovery check.
         let present_containers: std::collections::HashSet<String> = self
             .runtime
@@ -4667,6 +4676,27 @@ impl ContainerOrchestrator for ProdContainerOrchestrator {
         let lock = self.app_lock(app_id).await;
         let _guard = lock.lock().await;
         let name = compute_container_name(&lm.manifest);
+        let mut resolved = lm.manifest.clone();
+        resolve_catalog_image(&mut resolved);
+        if resolved.app.container.build.is_none() {
+            if let Some(target) = &resolved.app.container.image {
+                if let Ok(running) = self.runtime.get_container_status(&name).await {
+                    match crate::container::image_versions::compare_image_versions(
+                        target,
+                        &running.image,
+                    ) {
+                        Some(std::cmp::Ordering::Less) => anyhow::bail!(
+                            "Refusing to downgrade {} from {} to {} during update",
+                            app_id,
+                            running.image,
+                            target
+                        ),
+                        Some(std::cmp::Ordering::Equal) => return Ok(()),
+                        _ => {}
+                    }
+                }
+            }
+        }
         let _ = self.runtime.stop_container(&name).await;
         let _ = self.runtime.remove_container(&name).await;
         self.install_fresh(&lm).await
@@ -5076,6 +5106,7 @@ mod tests {
         calls: StdMutex<Vec<String>>,
         /// container_name -> ContainerState. Absence = "doesn't exist".
         containers: StdMutex<HashMap<String, ContainerState>>,
+        running_images: StdMutex<HashMap<String, String>>,
         /// container_name -> Podman health status.
         health: StdMutex<HashMap<String, String>>,
         /// image_ref -> present. Absence = "not present in local storage".
@@ -5200,7 +5231,13 @@ mod tests {
                 health,
                 exit_code: None,
                 started_at: None,
-                image: "test-image".to_string(),
+                image: self
+                    .running_images
+                    .lock()
+                    .unwrap()
+                    .get(name)
+                    .cloned()
+                    .unwrap_or_else(|| "test-image".to_string()),
                 created: "now".to_string(),
                 ports: vec![],
                 lan_address: None,
@@ -6370,6 +6407,42 @@ app:
         assert!(cascade_pairs_for_report(&r, &none).is_empty());
     }
 
+    #[tokio::test]
+    async fn reconcile_wallet_start_precedes_unrelated_failed_image_pull() {
+        let rt = Arc::new(MockRuntime::default());
+        rt.set_state("bitcoin-knots", ContainerState::Exited);
+        rt.set_state("lnd", ContainerState::Exited);
+        *rt.fail_pull.lock().unwrap() = Some("registry unreachable".into());
+        let mut orch = orch_with(rt.clone()).await;
+        orch.set_disk_gb_for_test(2000);
+        for id in ["unrelated", "lnd", "bitcoin-knots"] {
+            orch.insert_manifest_for_test(
+                pull_manifest(id, &format!("docker.io/example/{id}:1")),
+                PathBuf::from(format!("/tmp/{id}")),
+            )
+            .await;
+        }
+        let report = orch.reconcile_all().await;
+        assert!(report.failures.iter().any(|(id, _)| id == "unrelated"));
+        let calls = rt.calls();
+        let bitcoin = calls
+            .iter()
+            .position(|c| c == "start_container:bitcoin-knots")
+            .unwrap();
+        let lnd = calls
+            .iter()
+            .position(|c| c == "start_container:lnd")
+            .unwrap();
+        let pull = calls
+            .iter()
+            .position(|c| c.starts_with("pull_image:"))
+            .unwrap();
+        assert!(
+            bitcoin < lnd && lnd < pull,
+            "wallet startup was delayed by unrelated recovery: {calls:?}"
+        );
+    }
+
     #[tokio::test]
     async fn reconcile_starts_exited_container() {
         let rt = Arc::new(MockRuntime::default());
@@ -6771,6 +6844,41 @@ app:
         assert_eq!(ids, vec!["bitcoin-knots", "bitcoin-ui"]);
     }
 
+    #[tokio::test]
+    async fn upgrade_preserves_container_when_catalog_is_stale_or_already_installed() {
+        for (target, should_error) in [("v3.3.1", true), ("v3.3.1-archy1", false)] {
+            let rt = Arc::new(MockRuntime::default());
+            rt.set_state("update-regression", ContainerState::Running);
+            rt.running_images.lock().unwrap().insert(
+                "update-regression".into(),
+                "registry.test/old/mempool-frontend:v3.3.1-archy1".into(),
+            );
+            let orch = orch_with(rt.clone()).await;
+            orch.insert_manifest_for_test(
+                pull_manifest(
+                    "update-regression",
+                    &format!("registry.test/new/mempool-frontend:{target}"),
+                ),
+                PathBuf::from("/tmp/update-regression"),
+            )
+            .await;
+            assert_eq!(
+                orch.upgrade("update-regression").await.is_err(),
+                should_error
+            );
+            assert!(
+                !rt.calls()
+                    .iter()
+                    .any(|call| call.starts_with("stop_container:")
+                        || call.starts_with("remove_container:")
+                        || call.starts_with("pull_image:")
+                        || call.starts_with("create_container:")),
+                "{:?}",
+                rt.calls()
+            );
+        }
+    }
+
     #[tokio::test]
     async fn upgrade_removes_and_reinstalls() {
         let rt = Arc::new(MockRuntime::default());

core/archipelago/src/content_server.rs

@@ -5,13 +5,110 @@
 
 use anyhow::{Context, Result};
 use serde::{Deserialize, Serialize};
+use sha2::{Digest, Sha256};
+use std::collections::HashMap;
+use std::future::Future;
 use std::path::{Path, PathBuf};
+use std::sync::{Arc, LazyLock};
+use std::time::{Duration, Instant};
 use tokio::fs;
+use tokio::sync::Mutex;
 use tracing::{debug, warn};
 
 const CATALOG_FILE: &str = "content/catalog.json";
 const CONTENT_DIR: &str = "content/files";
 
+/// How long a redeemed payment token keeps entitling its buyer to re-fetch the
+/// item it paid for. Long enough to cover a buyer's transport fallback (FIPS →
+/// Tor re-sends the same request, token included) and a manual retry; short
+/// enough that the ledger stays tiny and a leaked token isn't a standing pass.
+const REDEMPTION_TTL: Duration = Duration::from_secs(600);
+
+/// One ledger slot per payment token (keyed by its SHA-256 — the raw bearer
+/// token is never held here). The inner mutex serialises verification of the
+/// same token; its value is the content id the token was redeemed for.
+struct RedemptionSlot {
+    created_at: Instant,
+    redeemed_for: Arc<Mutex<Option<String>>>,
+}
+
+static REDEMPTIONS: LazyLock<Mutex<HashMap<String, RedemptionSlot>>> =
+    LazyLock::new(|| Mutex::new(HashMap::new()));
+
+/// Decide whether `token` pays for `content_id`, redeeming it at most once.
+///
+/// Payment tokens are single-use: verifying one swaps its proofs at the mint,
+/// so a second verification of the same token always fails "already spent".
+/// A buyer's HTTP client can legitimately send the same request twice — its
+/// FIPS attempt gets a 404/5xx and it re-sends over Tor — and without this
+/// the seller redeemed the token on the first request, then answered the
+/// retry `402 Payment required`: money taken, file never delivered.
+///
+/// So the first verification that succeeds is remembered (per token, per
+/// item, for [`REDEMPTION_TTL`]) and later requests for the same item present
+/// the same token are authorised without touching the mint again. Concurrent
+/// requests with one token queue on the slot so only one runs `verify`.
+/// A failed verification is not remembered — the slot is dropped so garbage
+/// tokens can't accumulate and a legitimate retry gets a fresh attempt.
+async fn authorize_payment<F, Fut>(token: &str, content_id: &str, verify: F) -> bool
+where
+    F: FnOnce() -> Fut,
+    Fut: Future<Output = bool>,
+{
+    let key = hex::encode(Sha256::digest(token.as_bytes()));
+    let redeemed_for = {
+        let mut ledger = REDEMPTIONS.lock().await;
+        ledger.retain(|_, s| s.created_at.elapsed() < REDEMPTION_TTL);
+        ledger
+            .entry(key.clone())
+            .or_insert_with(|| RedemptionSlot {
+                created_at: Instant::now(),
+                redeemed_for: Arc::new(Mutex::new(None)),
+            })
+            .redeemed_for
+            .clone()
+    };
+
+    let mut state = redeemed_for.lock().await;
+    if state.as_deref() == Some(content_id) {
+        debug!(
+            "Payment token already redeemed for '{}' — serving without re-verifying",
+            content_id
+        );
+        return true;
+    }
+    if verify().await {
+        *state = Some(content_id.to_string());
+        return true;
+    }
+    // Keep a slot that already holds a redemption (this token paid for a
+    // different item); drop one that never verified anything.
+    let never_redeemed = state.is_none();
+    drop(state);
+    if never_redeemed {
+        REDEMPTIONS.lock().await.remove(&key);
+    }
+    false
+}
+
+/// Confirm the node can actually hand the file over: it exists and this
+/// process may read it. Must run BEFORE a payment is redeemed — a paid buyer
+/// who then hits a read error has lost their token for nothing (2026-09-18:
+/// filebrowser-owned `0640` files the node's service user couldn't open; the
+/// stat calls passed, `fs::read` failed after the swap, the buyer got a 404).
+/// Reading a byte (not just opening) also rejects a directory.
+async fn ensure_servable(file_path: &Path) -> Result<()> {
+    use tokio::io::AsyncReadExt;
+    let mut file = fs::File::open(file_path)
+        .await
+        .with_context(|| format!("content file {} is not readable", file_path.display()))?;
+    let mut probe = [0u8; 1];
+    file.read(&mut probe)
+        .await
+        .with_context(|| format!("content file {} cannot be read", file_path.display()))?;
+    Ok(())
+}
+
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct ContentItem {
     pub id: String,
@@ -296,6 +393,31 @@ pub async fn serve_content(
         }
     }
 
+    // Verify the file can be served BEFORE any payment is redeemed. The gate
+    // below swaps the buyer's token at the mint; failing to hand over the file
+    // after that takes their money and delivers nothing.
+    let file_path = content_file_path(data_dir, item);
+    if !file_path.exists() {
+        // The catalog entry survived (it's a separate JSON file) but its
+        // backing file is gone — most likely lost in an unrelated data-dir
+        // reset (a shared filebrowser file, 2026-07-01: two catalog entries
+        // outlived a filebrowser reinstall that wiped the files themselves).
+        // Leaving the entry in place would keep advertising it as available
+        // to every peer forever, each hitting the exact same dead end this
+        // one just did. Prune it so it stops being offered.
+        warn!(
+            content_id = %id,
+            filename = %item.filename,
+            "content catalog entry's file is missing on disk — pruning the stale entry"
+        );
+        prune_missing_content_entry(data_dir, id).await;
+        return Ok(ServeResult::NotFound);
+    }
+    if let Err(e) = ensure_servable(&file_path).await {
+        warn!(content_id = %id, "cannot serve content (payment not taken): {e:#}");
+        return Err(e);
+    }
+
     // Check access control
     if !owner_session {
         match &item.access {
@@ -309,7 +431,10 @@ pub async fn serve_content(
                 if let Some(token) = payment_token {
                     if (method_accepted(&item.access, "ecash")
                         || method_accepted(&item.access, "fedimint"))
-                        && verify_payment_token(data_dir, token, *price_sats).await
+                        && authorize_payment(token, id, || {
+                            verify_payment_token(data_dir, token, *price_sats)
+                        })
+                        .await
                     {
                         authorized = true;
                     }
@@ -336,24 +461,6 @@ pub async fn serve_content(
         }
     }
 
-    let file_path = content_file_path(data_dir, item);
-    if !file_path.exists() {
-        // The catalog entry survived (it's a separate JSON file) but its
-        // backing file is gone — most likely lost in an unrelated data-dir
-        // reset (a shared filebrowser file, 2026-07-01: two catalog entries
-        // outlived a filebrowser reinstall that wiped the files themselves).
-        // Leaving the entry in place would keep advertising it as available
-        // to every peer forever, each hitting the exact same dead end this
-        // one just did. Prune it so it stops being offered.
-        warn!(
-            content_id = %id,
-            filename = %item.filename,
-            "content catalog entry's file is missing on disk — pruning the stale entry"
-        );
-        prune_missing_content_entry(data_dir, id).await;
-        return Ok(ServeResult::NotFound);
-    }
-
     let metadata = fs::metadata(&file_path)
         .await
         .context("Failed to read file metadata")?;
@@ -725,3 +832,182 @@ mod prune_missing_content_tests {
         assert_eq!(reloaded.items[0].id, "present-item");
     }
 }
+
+#[cfg(test)]
+mod paid_delivery_tests {
+    use super::*;
+    use std::sync::atomic::{AtomicUsize, Ordering};
+
+    /// A verifier that counts how often it actually runs.
+    fn counting(
+        calls: &Arc<AtomicUsize>,
+        result: bool,
+    ) -> impl FnOnce() -> std::future::Ready<bool> {
+        let calls = calls.clone();
+        move || {
+            calls.fetch_add(1, Ordering::SeqCst);
+            std::future::ready(result)
+        }
+    }
+
+    #[tokio::test]
+    async fn replayed_token_is_served_without_redeeming_twice() {
+        // The 2026-09-18 incident: the buyer's client re-sent the same request
+        // over Tor after the seller had already redeemed the token, and the
+        // second verification ("already spent") turned into a 402.
+        let calls = Arc::new(AtomicUsize::new(0));
+        assert!(authorize_payment("tok-replay", "item-a", counting(&calls, true)).await);
+        assert!(authorize_payment("tok-replay", "item-a", counting(&calls, true)).await);
+        assert_eq!(calls.load(Ordering::SeqCst), 1, "mint must be hit once");
+    }
+
+    #[tokio::test]
+    async fn concurrent_requests_with_one_token_redeem_once() {
+        // FIPS attempt still in flight when the Tor fallback arrives.
+        let calls = Arc::new(AtomicUsize::new(0));
+        let slow = |calls: Arc<AtomicUsize>| {
+            move || async move {
+                calls.fetch_add(1, Ordering::SeqCst);
+                tokio::time::sleep(Duration::from_millis(100)).await;
+                true
+            }
+        };
+        let (a, b) = tokio::join!(
+            authorize_payment("tok-concurrent", "item-a", slow(calls.clone())),
+            authorize_payment("tok-concurrent", "item-a", slow(calls.clone())),
+        );
+        assert!(a && b, "both requests must be served");
+        assert_eq!(calls.load(Ordering::SeqCst), 1);
+    }
+
+    #[tokio::test]
+    async fn failed_verification_is_not_remembered() {
+        let calls = Arc::new(AtomicUsize::new(0));
+        assert!(!authorize_payment("tok-bad", "item-a", counting(&calls, false)).await);
+        // A retry gets a fresh attempt — and can succeed (e.g. mint was down).
+        assert!(authorize_payment("tok-bad", "item-a", counting(&calls, true)).await);
+        assert_eq!(calls.load(Ordering::SeqCst), 2);
+        let ledger = REDEMPTIONS.lock().await;
+        let key = hex::encode(Sha256::digest(b"tok-bad"));
+        assert!(ledger.contains_key(&key), "successful redemption is kept");
+    }
+
+    #[tokio::test]
+    async fn failed_verification_leaves_no_ledger_entry() {
+        let calls = Arc::new(AtomicUsize::new(0));
+        assert!(!authorize_payment("tok-garbage", "item-a", counting(&calls, false)).await);
+        let key = hex::encode(Sha256::digest(b"tok-garbage"));
+        assert!(
+            !REDEMPTIONS.lock().await.contains_key(&key),
+            "garbage tokens must not accumulate"
+        );
+    }
+
+    #[tokio::test]
+    async fn token_redeemed_for_one_item_does_not_unlock_another() {
+        let calls = Arc::new(AtomicUsize::new(0));
+        assert!(authorize_payment("tok-cross", "item-a", counting(&calls, true)).await);
+        // Item B is verified on its own merits (the real mint would say
+        // "already spent"); it must not ride on item A's redemption…
+        assert!(!authorize_payment("tok-cross", "item-b", counting(&calls, false)).await);
+        assert_eq!(calls.load(Ordering::SeqCst), 2);
+        // …and failing there must not revoke what the token already paid for.
+        assert!(authorize_payment("tok-cross", "item-a", counting(&calls, true)).await);
+        assert_eq!(calls.load(Ordering::SeqCst), 2);
+    }
+
+    fn paid_item(id: &str, filename: &str) -> ContentItem {
+        ContentItem {
+            id: id.to_string(),
+            filename: filename.to_string(),
+            mime_type: "audio/mpeg".to_string(),
+            size_bytes: 4,
+            description: String::new(),
+            access: AccessControl::Paid {
+                price_sats: 10,
+                accepted: vec!["ecash".to_string()],
+            },
+            availability: Availability::AllPeers,
+            added_at: "2026-01-01T00:00:00Z".to_string(),
+        }
+    }
+
+    #[cfg(unix)]
+    #[tokio::test]
+    async fn unreadable_paid_file_errors_before_any_payment_is_redeemed() {
+        // Filebrowser-owned 0640 files the node's service user can't read:
+        // stat() succeeds, read() fails. That must surface as an error BEFORE
+        // the token is verified — never after the swap has taken the money.
+        use std::os::unix::fs::PermissionsExt;
+        let dir = tempfile::tempdir().unwrap();
+        let data_dir = dir.path();
+        save_catalog(
+            data_dir,
+            &ContentCatalog {
+                items: vec![paid_item("locked", "locked.mp3")],
+            },
+        )
+        .await
+        .unwrap();
+        let files = data_dir.join("content").join("files");
+        tokio::fs::create_dir_all(&files).await.unwrap();
+        let file = files.join("locked.mp3");
+        tokio::fs::write(&file, b"data").await.unwrap();
+        std::fs::set_permissions(&file, std::fs::Permissions::from_mode(0o000)).unwrap();
+        if std::fs::File::open(&file).is_ok() {
+            return; // running as root: permissions can't be enforced here
+        }
+
+        // A token that would fail verification if it were reached: getting
+        // PaymentRequired here would mean the gate ran before the file check.
+        let result = serve_content(
+            data_dir,
+            "locked",
+            Some("cashuBnot-a-real-token"),
+            None,
+            None,
+            None,
+            false,
+        )
+        .await;
+        assert!(
+            result.is_err(),
+            "unreadable file must be a server error, not 402/404"
+        );
+        let key = hex::encode(Sha256::digest(b"cashuBnot-a-real-token"));
+        assert!(
+            !REDEMPTIONS.lock().await.contains_key(&key),
+            "no redemption may be attempted for an unservable file"
+        );
+    }
+
+    #[tokio::test]
+    async fn readable_paid_file_with_bad_token_still_requires_payment() {
+        let dir = tempfile::tempdir().unwrap();
+        let data_dir = dir.path();
+        save_catalog(
+            data_dir,
+            &ContentCatalog {
+                items: vec![paid_item("ok", "ok.mp3")],
+            },
+        )
+        .await
+        .unwrap();
+        let files = data_dir.join("content").join("files");
+        tokio::fs::create_dir_all(&files).await.unwrap();
+        tokio::fs::write(files.join("ok.mp3"), b"data").await.unwrap();
+
+        let result = serve_content(
+            data_dir,
+            "ok",
+            Some("cashuBnot-a-real-token-2"),
+            None,
+            None,
+            None,
+            false,
+        )
+        .await
+        .unwrap();
+        assert!(matches!(result, ServeResult::PaymentRequired(10)));
+    }
+}

core/archipelago/src/update.rs

@@ -2159,20 +2159,25 @@ async fn apply_per_app_auto_updates(
     }
 }
 
-/// After a catalog refresh that changed the cached bytes, rebuild the
-/// orchestrator's manifest map so registry-shipped manifest changes take
-/// effect now instead of at the next service restart.
-async fn reload_manifests_if_changed(
-    refresh: crate::container::app_catalog::CatalogRefresh,
+/// Reload after every successful refresh, including unchanged bytes: the cache
+/// may have been written before a previous reload failed. Auto-updates only run
+/// when the catalog and the orchestrator's manifests are ready together.
+async fn reload_catalog_manifests(
+    _refresh: crate::container::app_catalog::CatalogRefresh,
     orchestrator: &Option<std::sync::Arc<dyn crate::container::traits::ContainerOrchestrator>>,
-) {
-    if !refresh.changed {
-        return;
-    }
-    let Some(orch) = orchestrator else { return };
+) -> bool {
+    let Some(orch) = orchestrator else {
+        return false;
+    };
     match orch.reload_manifests().await {
-        Ok(n) => info!("Update scheduler: catalog changed, reloaded {n} manifest(s)"),
-        Err(e) => warn!("Update scheduler: manifest reload after catalog change failed: {e}"),
+        Ok(n) => {
+            info!("Update scheduler: refreshed catalog, reloaded {n} manifest(s)");
+            true
+        }
+        Err(e) => {
+            warn!("Update scheduler: manifest reload failed; skipping auto-updates: {e}");
+            false
+        }
     }
 }
 
@@ -2188,7 +2193,9 @@ pub async fn run_update_scheduler(
     // Refresh the app catalog once at startup so per-app "update available"
     // badges appear without waiting for the first hourly tick.
     match crate::container::app_catalog::refresh_catalog(&data_dir).await {
-        Ok(refresh) => reload_manifests_if_changed(refresh, &orchestrator).await,
+        Ok(refresh) => {
+            reload_catalog_manifests(refresh, &orchestrator).await;
+        }
         Err(e) => debug!(
             "Update scheduler: initial app-catalog refresh failed: {}",
             e
@@ -2204,14 +2211,22 @@ pub async fn run_update_scheduler(
         // previously cached catalog stays in place (origin-always-wins).
         // A changed catalog also reloads the orchestrator's manifest overlay so
         // catalog-shipped manifest fixes apply without a service restart.
-        match crate::container::app_catalog::refresh_catalog(&data_dir).await {
-            Ok(refresh) => reload_manifests_if_changed(refresh, &orchestrator).await,
-            Err(e) => debug!("Update scheduler: app-catalog refresh failed: {}", e),
-        }
+        let catalog_ready = match crate::container::app_catalog::refresh_catalog(&data_dir).await {
+            Ok(refresh) => reload_catalog_manifests(refresh, &orchestrator).await,
+            Err(e) => {
+                debug!(
+                    "Update scheduler: app-catalog refresh failed; skipping auto-updates: {}",
+                    e
+                );
+                false
+            }
+        };
 
-        // Per-app auto-update-to-latest (multi-version support). Runs every tick
-        // regardless of the binary-OTA schedule below; opt-in + pin-respecting.
-        apply_per_app_auto_updates(&orchestrator).await;
+        // Per-app updates require fresh, loaded manifests; a failed refresh
+        // may still show cached badges but must not trigger container changes.
+        if catalog_ready {
+            apply_per_app_auto_updates(&orchestrator).await;
+        }
 
         let state = match load_state(&data_dir).await {
             Ok(s) => s,

core/archipelago/src/wallet/ecash.rs

@@ -1205,6 +1205,7 @@ pub async fn receive_token(data_dir: &Path, token_str: &str) -> Result<u64> {
     // for the log. Remember the last one so a total failure can tell the user
     // *why* instead of just "nothing was received".
     let mut last_reason: Option<String> = None;
+    let mut all_already_redeemed = true;
 
     // Swap proofs at each mint
     for entry in &token.token {
@@ -1217,6 +1218,7 @@ pub async fn receive_token(data_dir: &Path, token_str: &str) -> Result<u64> {
             }
             Err(e) => {
                 warn!("Failed to swap proofs from mint {}: {:#}", entry.mint, e);
+                all_already_redeemed &= e.is::<super::mint_client::AlreadyRedeemed>();
                 last_reason = Some(e.to_string());
                 // Continue with other mints if any
             }
@@ -1224,10 +1226,7 @@ pub async fn receive_token(data_dir: &Path, token_str: &str) -> Result<u64> {
     }
 
     if received_total == 0 {
-        match last_reason {
-            Some(reason) => anyhow::bail!("Could not receive this ecash: {}", reason),
-            None => anyhow::bail!("Failed to receive any proofs from token"),
-        }
+        return Err(receive_failure(last_reason, all_already_redeemed));
     }
 
     wallet.record_tx(
@@ -1243,6 +1242,17 @@ pub async fn receive_token(data_dir: &Path, token_str: &str) -> Result<u64> {
     Ok(received_total)
 }
 
+fn receive_failure(last_reason: Option<String>, all_already_redeemed: bool) -> anyhow::Error {
+    match last_reason {
+        Some(reason) if all_already_redeemed => {
+            anyhow::Error::new(super::mint_client::AlreadyRedeemed)
+                .context(format!("Could not receive this ecash: {reason}"))
+        }
+        Some(reason) => anyhow::anyhow!("Could not receive this ecash: {reason}"),
+        None => anyhow::anyhow!("Failed to receive any proofs from token"),
+    }
+}
+
 /// Receive a legacy format token (cashuSend_{amount}_{uuid}_{timestamp}).
 /// For backwards compatibility during migration period.
 async fn receive_legacy_token(data_dir: &Path, token_str: &str) -> Result<u64> {
@@ -1632,6 +1642,18 @@ fn default_mint_url() -> String {
 
 #[cfg(test)]
 mod tests {
+    #[test]
+    fn mixed_mint_failures_do_not_discard_a_retryable_claim() {
+        let reason = super::super::mint_client::ALREADY_REDEEMED_MSG.to_string();
+        assert!(super::receive_failure(Some(reason.clone()), true)
+            .is::<super::super::mint_client::AlreadyRedeemed>());
+        assert!(!super::receive_failure(Some(reason), false)
+            .is::<super::super::mint_client::AlreadyRedeemed>());
+        assert!(
+            !super::receive_failure(None, true).is::<super::super::mint_client::AlreadyRedeemed>()
+        );
+    }
+
     use super::*;
     use tempfile::TempDir;
 

core/archipelago/src/wallet/minibits.rs

@@ -47,7 +47,8 @@
 //! key, a crash mid-loop) must not silently lose the coins, so every fetched
 //! token is persisted to `MinibitsState::pending_claims` *before* decrypt/
 //! redeem is attempted, and stays there — retried on every later poll — until
-//! it succeeds. `ClaimOutcome::failed_count` reports how many are still
+//! it succeeds or every mint reports that it was already spent.
+//! `ClaimOutcome::failed_count` reports how many are still
 //! stuck so the caller can surface it instead of it being a log-only event.
 //! Separately, `ensure_mint_accepted` keeps the Minibits mint on the node's
 //! accepted-mints allow-list: the address is inherently backed by that one
@@ -167,6 +168,9 @@ pub struct MinibitsState {
     /// already-spent token) but wasteful and noisy.
     #[serde(default)]
     pub last_dm_seen_at: u64,
+    /// Resume a bounded backward scan before advancing to newer relay events.
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub relay_scan: Option<RelayScan>,
     /// Event ids already queued from the relay. `created_at` has only
     /// one-second resolution, so a strict `since = last + 1` watermark can
     /// permanently miss a second payment published later in the same second.
@@ -627,6 +631,7 @@ async fn register_new_state(
         created_at: chrono::Utc::now().to_rfc3339(),
         pending_claims: Vec::new(),
         last_dm_seen_at: 0,
+        relay_scan: None,
         seen_dm_ids: Vec::new(),
         last_receipt_id: 0,
         last_receipt_sats: 0,
@@ -652,6 +657,16 @@ pub struct ClaimOutcome {
     pub receipt_at: u64,
 }
 
+/// True when `ecash::receive_token` failed because the token was already
+/// redeemed (mint error 11001, see `mint_client::describe_mint_error_code`) —
+/// a terminal condition, not a reason to retry. Seen on a deployed node,
+/// 2026-09-15: a claim that had already been swept kept failing this way on
+/// every poll forever, since nothing distinguished it from a transient
+/// failure worth retrying.
+fn is_already_redeemed(err: &anyhow::Error) -> bool {
+    err.is::<super::mint_client::AlreadyRedeemed>()
+}
+
 const NO_CLAIMS: ClaimOutcome = ClaimOutcome {
     claimed_count: 0,
     received_sats: 0,
@@ -697,38 +712,58 @@ fn outcome_with_latest_receipt(
 /// three real payments that `/claim` never surfaced. Best-effort: a relay
 /// error here must not abort the poll, since `pending_claims` may still hold
 /// earlier fetches worth retrying.
+///
+/// Queries `RELAY_URL` (the service's own relay) alone first — the happy
+/// path for a poll is one WebSocket connection, not three, and the wallet's
+/// derived Nostr pubkey isn't broadcast to the public fallback relays unless
+/// it's actually needed. Only when that relay is unreachable does it fall
+/// back to all of `CLAIM_RELAY_URLS`. Results are paged (capped at
+/// `CLAIM_MAX_PAGES`) since a relay returns only the newest `limit` events for
+/// a filter. A durable backward cursor keeps older pages reachable even after
+/// newly queued claims advance the normal forward watermark.
 async fn fetch_relay_dms(
     our_pubkey: nostr_sdk::PublicKey,
     server_pubkey: nostr_sdk::PublicKey,
     since: u64,
-) -> Vec<(String, u64, String, String)> {
+    resume: Option<RelayScan>,
+) -> RelayBatch {
     let client = Client::default();
-    for url in CLAIM_RELAY_URLS {
-        if let Err(e) = client.add_relay(*url).await {
-            warn!("Minibits: could not add relay {url}: {e}");
+    if let Err(e) = client.add_relay(RELAY_URL).await {
+        warn!("Minibits: could not add relay {RELAY_URL}: {e}");
+    }
+    let primary_reachable = client
+        .try_connect_relay(RELAY_URL, std::time::Duration::from_secs(3))
+        .await
+        .is_ok();
+    if !primary_reachable {
+        warn!("Minibits: primary relay {RELAY_URL} unreachable, falling back to public relays too");
+        for url in &CLAIM_RELAY_URLS[1..] {
+            if let Err(e) = client.add_relay(*url).await {
+                warn!("Minibits: could not add relay {url}: {e}");
+            }
         }
+        client.connect().await;
     }
-    client.connect().await;
     // Give relays a moment to finish the WebSocket handshake before the
     // fetch's own timeout starts consuming that time.
     tokio::time::sleep(std::time::Duration::from_millis(400)).await;
 
-    // Nostr timestamps have one-second resolution. Query the boundary second
-    // inclusively: a later-published payment may legitimately share that
-    // timestamp. `seen_dm_ids` performs the exact deduplication locally.
-    let filter = Filter::new()
-        .author(server_pubkey)
-        .pubkey(our_pubkey)
-        .kind(Kind::from(4u16))
-        .since(Timestamp::from(since))
-        .limit(200);
-
-    let result = match client
-        .fetch_events(filter, std::time::Duration::from_secs(5))
-        .await
-    {
-        Ok(events) => {
-            let mut out: Vec<(String, u64, String, String)> = events
+    let batch = collect_relay_pages(since, resume, |scan| {
+        let client = &client;
+        async move {
+            let mut filter = Filter::new()
+                .author(server_pubkey)
+                .pubkey(our_pubkey)
+                .kind(Kind::from(4u16))
+                .since(Timestamp::from(scan.since))
+                .limit(scan.limit);
+            if let Some(until) = scan.until {
+                filter = filter.until(Timestamp::from(until));
+            }
+            let events = client
+                .fetch_events(filter, std::time::Duration::from_secs(5))
+                .await?;
+            Ok(events
                 .into_iter()
                 .map(|e| {
                     (
@@ -738,18 +773,83 @@ async fn fetch_relay_dms(
                         e.id.to_hex(),
                     )
                 })
-                .collect();
-            out.sort_by_key(|(_, created_at, _, _)| *created_at);
-            out
+                .collect())
         }
-        Err(e) => {
-            warn!("Minibits: relay fetch for claim DMs failed: {e}");
-            Vec::new()
-        }
-    };
-
+    })
+    .await;
     client.shutdown().await;
-    result
+    batch
+}
+
+const CLAIM_PAGE_LIMIT: usize = 200;
+const CLAIM_MAX_PAGES: usize = 5;
+type RelayDm = (String, u64, String, String);
+
+#[derive(Debug, Clone, Copy, Serialize, Deserialize)]
+pub struct RelayScan {
+    since: u64,
+    until: Option<u64>,
+    limit: usize,
+}
+
+struct RelayBatch {
+    dms: Vec<RelayDm>,
+    resume: Option<RelayScan>,
+}
+
+/// NIP-01 returns newest events first. Walk backward with an inclusive `until`
+/// boundary, deduplicating event ids. A full boundary second needs a larger
+/// limit, not `until - 1`, which would skip payments sharing that timestamp.
+/// Persist the cursor at the page cap or on failure so older claims cannot be
+/// hidden by the newest timestamp already queued in `last_dm_seen_at`.
+async fn collect_relay_pages<F, Fut>(
+    since: u64,
+    resume: Option<RelayScan>,
+    mut fetch: F,
+) -> RelayBatch
+where
+    F: FnMut(RelayScan) -> Fut,
+    Fut: std::future::Future<Output = Result<Vec<RelayDm>>>,
+{
+    let mut scan = resume.unwrap_or(RelayScan {
+        since,
+        until: None,
+        limit: CLAIM_PAGE_LIMIT,
+    });
+    let mut out = Vec::new();
+    let mut ids = std::collections::HashSet::new();
+    let mut resume = Some(scan);
+    for _ in 0..CLAIM_MAX_PAGES {
+        let events = match fetch(scan).await {
+            Ok(events) => events,
+            Err(e) => {
+                warn!("Minibits: relay fetch failed; preserving scan cursor: {e}");
+                break;
+            }
+        };
+        let count = events.len();
+        let oldest = events.iter().map(|e| e.1).min();
+        for event in events {
+            if ids.insert(event.3.clone()) {
+                out.push(event);
+            }
+        }
+        if count < scan.limit {
+            resume = None;
+            break;
+        }
+        if let Some(oldest) = oldest {
+            if scan.until == Some(oldest) {
+                scan.limit = scan.limit.saturating_add(CLAIM_PAGE_LIMIT);
+            } else {
+                scan.until = Some(oldest);
+                scan.limit = CLAIM_PAGE_LIMIT;
+            }
+        }
+        resume = Some(scan);
+    }
+    out.sort_by(|a, b| (a.1, &a.3).cmp(&(b.1, &b.3)));
+    RelayBatch { dms: out, resume }
 }
 
 fn queue_relay_dm(
@@ -907,8 +1007,15 @@ pub async fn claim_and_redeem(data_dir: &Path) -> Result<ClaimOutcome> {
     // NIP-04 DM on relays, not via `/claim` above. `since` is our own
     // watermark (Nostr events never expire off a relay, so without it we'd
     // re-fetch and re-attempt every claim ever sent on every poll).
-    let dms = fetch_relay_dms(identity.keys.public_key(), server_pk, state.last_dm_seen_at).await;
-    for (content, created_at, author, event_id) in dms {
+    let batch = fetch_relay_dms(
+        identity.keys.public_key(),
+        server_pk,
+        state.last_dm_seen_at,
+        state.relay_scan,
+    )
+    .await;
+    state.relay_scan = batch.resume;
+    for (content, created_at, author, event_id) in batch.dms {
         if author != state.server_nostr_pubkey {
             warn!("Minibits: ignoring claim DM from unexpected pubkey {author}");
             continue;
@@ -964,6 +1071,14 @@ pub async fn claim_and_redeem(data_dir: &Path) -> Result<ClaimOutcome> {
                 sats += got;
                 info!("Minibits: redeemed a claimed payment ({got} sats)");
             }
+            Err(e) if is_already_redeemed(&e) => {
+                // Terminal: the value was already swept (a relay-watermark
+                // replay, or a claim redeemed by an earlier run before a
+                // crash lost track of it). Retrying can never succeed, so
+                // drop it instead of leaving `failed_count` stuck non-zero
+                // forever — see archy-x250-pa3, 2026-09-15.
+                info!("Minibits mint reports this claim was already redeemed; removing it from the retry queue");
+            }
             Err(e) => {
                 warn!("Minibits claim decrypted but failed to redeem ({e}); will retry next poll");
                 still_pending.push(claim.clone());
@@ -994,6 +1109,123 @@ pub async fn claim_and_redeem(data_dir: &Path) -> Result<ClaimOutcome> {
 
 #[cfg(test)]
 mod tests {
+    fn simulated_relay_page(events: &[RelayDm], scan: RelayScan) -> Vec<RelayDm> {
+        let mut page: Vec<_> = events
+            .iter()
+            .filter(|e| e.1 >= scan.since && scan.until.is_none_or(|until| e.1 <= until))
+            .cloned()
+            .collect();
+        page.sort_by(|a, b| b.1.cmp(&a.1).then_with(|| a.3.cmp(&b.3)));
+        page.truncate(scan.limit);
+        page
+    }
+
+    fn relay_fixture(count: usize, same_second: bool) -> Vec<RelayDm> {
+        (1..=count)
+            .map(|n| {
+                (
+                    format!("claim-{n}"),
+                    if same_second { 100 } else { n as u64 },
+                    "service".into(),
+                    format!("id-{n:06}"),
+                )
+            })
+            .collect()
+    }
+
+    #[tokio::test]
+    async fn relay_paging_fetches_older_claims_in_newest_first_backlog() {
+        let events = relay_fixture(450, false);
+        let batch = collect_relay_pages(0, None, |scan| {
+            std::future::ready(Ok(simulated_relay_page(&events, scan)))
+        })
+        .await;
+        assert_eq!(batch.dms.len(), 450);
+        assert!(batch.resume.is_none());
+        assert_eq!(batch.dms.first().unwrap().1, 1);
+        assert_eq!(batch.dms.last().unwrap().1, 450);
+    }
+
+    #[tokio::test]
+    async fn relay_paging_preserves_payments_at_the_same_timestamp() {
+        let events = relay_fixture(250, true);
+        let batch = collect_relay_pages(100, None, |scan| {
+            std::future::ready(Ok(simulated_relay_page(&events, scan)))
+        })
+        .await;
+        assert_eq!(batch.dms.len(), 250);
+        assert!(batch.resume.is_none());
+    }
+
+    #[tokio::test]
+    async fn relay_page_cap_resumes_older_claims_after_watermark_advances() {
+        let events = relay_fixture(1300, false);
+        let mut state = MinibitsState::default();
+        let first = collect_relay_pages(0, None, |scan| {
+            std::future::ready(Ok(simulated_relay_page(&events, scan)))
+        })
+        .await;
+        assert!(first.resume.is_some());
+        state.relay_scan = first.resume;
+        let mut ids = std::collections::HashSet::new();
+        for (content, time, author, id) in first.dms {
+            ids.insert(id.clone());
+            queue_relay_dm(&mut state, content, time, id, author);
+        }
+        assert_eq!(state.last_dm_seen_at, 1300);
+        let state: MinibitsState =
+            serde_json::from_str(&serde_json::to_string(&state).unwrap()).unwrap();
+        let second = collect_relay_pages(state.last_dm_seen_at, state.relay_scan, |scan| {
+            std::future::ready(Ok(simulated_relay_page(&events, scan)))
+        })
+        .await;
+        assert!(second.resume.is_none());
+        ids.extend(second.dms.into_iter().map(|e| e.3));
+        assert_eq!(ids.len(), 1300);
+    }
+
+    #[tokio::test]
+    async fn relay_fetch_failure_keeps_the_unfinished_page_cursor() {
+        let events = relay_fixture(450, false);
+        let mut requests = 0;
+        let first = collect_relay_pages(0, None, |scan| {
+            requests += 1;
+            std::future::ready(if requests == 1 {
+                Ok(simulated_relay_page(&events, scan))
+            } else {
+                Err(anyhow!("relay timeout"))
+            })
+        })
+        .await;
+        assert_eq!(first.dms.len(), 200);
+        assert_eq!(first.resume.unwrap().until, Some(251));
+        let second = collect_relay_pages(450, first.resume, |scan| {
+            std::future::ready(Ok(simulated_relay_page(&events, scan)))
+        })
+        .await;
+        let ids: std::collections::HashSet<_> = first
+            .dms
+            .into_iter()
+            .chain(second.dms)
+            .map(|e| e.3)
+            .collect();
+        assert_eq!(ids.len(), 450);
+    }
+
+    #[test]
+    fn only_typed_spent_claims_are_terminal_even_with_wrapped_errors() {
+        let spent = anyhow::Error::new(super::super::mint_client::AlreadyRedeemed)
+            .context("receive token")
+            .context("claim failed");
+        assert!(is_already_redeemed(&spent));
+        assert!(!is_already_redeemed(&anyhow!(
+            super::super::mint_client::ALREADY_REDEEMED_MSG
+        )));
+        assert!(!is_already_redeemed(&anyhow!(
+            "mint temporarily unreachable"
+        )));
+    }
+
     use super::*;
 
     #[test]

core/archipelago/src/wallet/mint_client.rs

@@ -71,10 +71,28 @@ pub struct MintResult {
 /// keyset codes shared by NUT-02/03/04/05 — the codes a swap/melt/mint call
 /// can actually hit. Returns `None` for anything else (e.g. Lightning/quote
 /// codes in the 20000s) so the caller falls back to the mint's own `detail`.
+///
+/// Text of the NUT error-code-11001 translation, exposed so callers that
+/// received an `anyhow::Error` from a receive/redeem path (e.g. a replayed
+/// Minibits claim) can recognize an already-spent token as terminal rather
+/// than retrying it forever.
+pub const ALREADY_REDEEMED_MSG: &str =
+    "This ecash has already been redeemed — it can't be claimed twice.";
+
+/// Typed terminal condition: never infer spent proofs from a mint's free text.
+#[derive(Debug)]
+pub(super) struct AlreadyRedeemed;
+impl std::fmt::Display for AlreadyRedeemed {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.write_str(ALREADY_REDEEMED_MSG)
+    }
+}
+impl std::error::Error for AlreadyRedeemed {}
+
 fn describe_mint_error_code(code: i64) -> Option<&'static str> {
     Some(match code {
         10001 => "The mint rejected these coins as invalid.",
-        11001 => "This ecash has already been redeemed — it can't be claimed twice.",
+        11001 => ALREADY_REDEEMED_MSG,
         11002 => "This ecash is already being redeemed elsewhere — try again in a moment.",
         11003 => "The mint already issued new coins for this exact request — there's nothing left to redeem.",
         11004 => "This request is still being processed by the mint — try again in a moment.",
@@ -124,8 +142,15 @@ fn describe_mint_error_body(status: reqwest::StatusCode, body: &str) -> String {
 /// translation layered on top via `.context()` so `{}` — what reaches the
 /// wallet user — shows something actionable instead of raw mint JSON.
 fn mint_error(op: &str, status: reqwest::StatusCode, body: &str) -> anyhow::Error {
-    let friendly = describe_mint_error_body(status, body);
-    anyhow::anyhow!("{} failed ({}): {}", op, status, body).context(friendly)
+    let cause = anyhow::anyhow!("{} failed ({}): {}", op, status, body);
+    if serde_json::from_str::<serde_json::Value>(body)
+        .ok()
+        .and_then(|v| v.get("code").and_then(|c| c.as_i64()))
+        == Some(11001)
+    {
+        return cause.context(AlreadyRedeemed);
+    }
+    cause.context(describe_mint_error_body(status, body))
 }
 
 /// HTTP client for a single Cashu mint.
@@ -803,6 +828,28 @@ impl MintClient {
 
 #[cfg(test)]
 mod tests {
+    #[test]
+    fn spent_condition_comes_from_code_not_remote_text_and_survives_context() {
+        let spent = super::mint_error(
+            "Swap",
+            reqwest::StatusCode::BAD_REQUEST,
+            r#"{"code":11001,"detail":"Token Already Spent"}"#,
+        )
+        .context("Receive failed");
+        assert!(spent.is::<super::AlreadyRedeemed>());
+        let body =
+            serde_json::json!({"code":11002,"detail":super::ALREADY_REDEEMED_MSG}).to_string();
+        assert!(
+            !super::mint_error("Swap", reqwest::StatusCode::BAD_REQUEST, &body)
+                .is::<super::AlreadyRedeemed>()
+        );
+        let body = serde_json::json!({"detail":super::ALREADY_REDEEMED_MSG}).to_string();
+        assert!(
+            !super::mint_error("Swap", reqwest::StatusCode::BAD_GATEWAY, &body)
+                .is::<super::AlreadyRedeemed>()
+        );
+    }
+
     use super::*;
 
     #[test]

docker/mempool-frontend/Dockerfile

@@ -1,14 +1,13 @@
 # Archipelago mempool frontend — adds a resilient nginx backend proxy.
 #
-# The only delta vs the upstream image is /patch/entrypoint.sh, which rewrites
-# the generated nginx-mempool.conf to use `resolver` + a variable proxy_pass so
-# the frontend re-resolves the backend (mempool-api) via DNS on every request.
-# Without this, nginx pins the backend IP at startup and serves 502 / "offline"
-# after any backend restart (podman reassigns the IP). See the script header.
-ARG BASE=source.archipelago-foundation.org/lfg2025/mempool-frontend:v3.0.0
+# Keep the upstream startup logic; repair its rendered proxy configuration.
+# Publish this derived image under an Archipelago-specific tag, never the
+# upstream version tag that the registry mirror can overwrite.
+ARG BASE=source.archipelago-foundation.org/lfg2025/mempool-frontend@sha256:d63498a109622475c913db4e3199d893f2440a451450e542923d2e55a38407a0
 FROM ${BASE}
 
 # --chmod keeps the exec bit (build runs as USER 1000, plain COPY lands root:0644
 # → "not executable"). Base USER/ENTRYPOINT/CMD (1000 / /patch/entrypoint.sh /
 # nginx -g "daemon off;") are inherited unchanged.
-COPY --chmod=0755 entrypoint.sh /patch/entrypoint.sh
+RUN cp /patch/entrypoint.sh /patch/upstream-entrypoint.sh
+COPY --chmod=0755 entrypoint.sh start-nginx.sh repair-nginx.sh /patch/

docker/mempool-frontend/README.md

@@ -0,0 +1,29 @@
+# Mempool frontend DNS recovery
+
+The stock v3.3.1 nginx configuration resolves `mempool-api` only when workers
+start. Recreating the backend can change its Podman address while the frontend
+continues to serve its static page, leaving all API/WebSocket requests offline.
+
+Build and test the derived image before publishing:
+
+```sh
+podman build --pull=never -t source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1 docker/mempool-frontend
+python3 scripts/test-mempool-dns-recovery.py
+```
+
+The base is pinned by digest. The wrapper preserves upstream runtime options,
+then repairs all four local API/WebSocket routes after placeholder rendering.
+DNS is cached for five seconds using the container network resolver. Explicit
+rewrites preserve API prefixes and query arguments; backend absence does not
+prevent nginx startup. An unexpected upstream configuration fails startup
+instead of silently omitting the fix.
+
+Use an Archipelago-specific image tag. Do not replace it with a stock upstream
+mirror when updating mempool. Every upstream update must rebuild this wrapper
+and pass the recovery test (backend absent, changed IP, HTTP and WebSocket
+mapping, repeated repair, and frontend restart).
+
+Publish the tested image before publishing the signed app catalog. Both the
+mempool umbrella image mapping and the archy-mempool-web embedded manifest must
+point at the patched image. Keep scripts/image-versions.sh in sync. The frontend
+health check must reach `/api/v1/backend-info` through nginx, not only `/`.

docker/mempool-frontend/entrypoint.sh

@@ -1,137 +1,5 @@
 #!/bin/sh
-__MEMPOOL_BACKEND_MAINNET_HTTP_HOST__=${BACKEND_MAINNET_HTTP_HOST:=127.0.0.1}
-__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__=${BACKEND_MAINNET_HTTP_PORT:=8999}
-__MEMPOOL_FRONTEND_HTTP_PORT__=${FRONTEND_HTTP_PORT:=8080}
-
-CONF=/etc/nginx/conf.d/nginx-mempool.conf
-
-# ─── archipelago patch ────────────────────────────────────────────────────
-# The stock frontend writes `proxy_pass http://<backend>:8999` with a literal
-# hostname and NO resolver, so nginx resolves the backend IP ONCE at worker
-# start and caches it for the process lifetime. Podman reassigns the backend
-# container's IP whenever it is restarted/recreated (gate, OTA, crash, reboot
-# re-IPAM), after which nginx keeps proxying to the dead IP → /api hangs, the
-# websocket 502s, and the mempool UI shows "offline" until nginx is reloaded.
-#
-# Fix: force per-request DNS re-resolution via `resolver` + a variable in
-# proxy_pass. Because a variable in proxy_pass disables nginx's automatic
-# location→URI rewriting, each block is rewritten to preserve its original
-# path mapping exactly:
-#   /api/v1/ws, /ws → "/"            (var + "/" replaces the whole URI)
-#   /api/v1         → identity       (no-URI proxy_pass passes $uri unchanged)
-#   /api/           → /api/v1/$1     (explicit rewrite, then no-URI proxy_pass)
-# Operates on the __PLACEHOLDER__ tokens so the host/port sed below fills in
-# the concrete values (incl. the `set $mp_backend` line). Idempotent.
-# Resolver address: podman's aardvark-dns answers on the network gateway
-# (e.g. 10.89.0.1), NOT Docker's 127.0.0.11. Read it from resolv.conf so this
-# works on any podman network/subnet (and still falls back for Docker).
-ARCHY_RESOLVER=$(awk '/^nameserver/ { print $2; exit }' /etc/resolv.conf 2>/dev/null)
-ARCHY_RESOLVER=${ARCHY_RESOLVER:-127.0.0.11}
-
-if ! grep -q 'set \$mp_backend' "$CONF"; then
-  awk -v res_addr="$ARCHY_RESOLVER" '
-    BEGIN { res = 0 }
-    /^[[:space:]]*location / && res == 0 {
-      print "\tresolver " res_addr " valid=10s ipv6=off;"
-      res = 1
-    }
-    /proxy_pass http:\/\/__MEMPOOL_BACKEND_MAINNET_HTTP_HOST__:__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__\/;/ {
-      print "\t\tset $mp_backend __MEMPOOL_BACKEND_MAINNET_HTTP_HOST__;"
-      print "\t\tproxy_pass http://$mp_backend:__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__/;"
-      next
-    }
-    /proxy_pass http:\/\/__MEMPOOL_BACKEND_MAINNET_HTTP_HOST__:__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__\/api\/v1\/;/ {
-      print "\t\tset $mp_backend __MEMPOOL_BACKEND_MAINNET_HTTP_HOST__;"
-      print "\t\trewrite ^/api/(.*)$ /api/v1/$1 break;"
-      print "\t\tproxy_pass http://$mp_backend:__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__;"
-      next
-    }
-    /proxy_pass http:\/\/__MEMPOOL_BACKEND_MAINNET_HTTP_HOST__:__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__\/api\/v1;/ {
-      print "\t\tset $mp_backend __MEMPOOL_BACKEND_MAINNET_HTTP_HOST__;"
-      print "\t\tproxy_pass http://$mp_backend:__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__;"
-      next
-    }
-    { print }
-  ' "$CONF" > "$CONF.archy" && mv "$CONF.archy" "$CONF"
-fi
-# ─── end archipelago patch ────────────────────────────────────────────────
-
-sed -i "s/__MEMPOOL_BACKEND_MAINNET_HTTP_HOST__/${__MEMPOOL_BACKEND_MAINNET_HTTP_HOST__}/g" /etc/nginx/conf.d/nginx-mempool.conf
-sed -i "s/__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__/${__MEMPOOL_BACKEND_MAINNET_HTTP_PORT__}/g" /etc/nginx/conf.d/nginx-mempool.conf
-
-cp /etc/nginx/nginx.conf /patch/nginx.conf
-sed -i "s/__MEMPOOL_FRONTEND_HTTP_PORT__/${__MEMPOOL_FRONTEND_HTTP_PORT__}/g" /patch/nginx.conf
-cat /patch/nginx.conf > /etc/nginx/nginx.conf
-
-if [ "${LIGHTNING_DETECTED_PORT}" != "" ];then
-  export LIGHTNING=true
-fi
-
-# Runtime overrides - read env vars defined in docker compose
-
-__MAINNET_ENABLED__=${MAINNET_ENABLED:=true}
-__TESTNET_ENABLED__=${TESTNET_ENABLED:=false}
-__TESTNET4_ENABLED__=${TESTNET_ENABLED:=false}
-__SIGNET_ENABLED__=${SIGNET_ENABLED:=false}
-__LIQUID_ENABLED__=${LIQUID_ENABLED:=false}
-__LIQUID_TESTNET_ENABLED__=${LIQUID_TESTNET_ENABLED:=false}
-__ITEMS_PER_PAGE__=${ITEMS_PER_PAGE:=10}
-__KEEP_BLOCKS_AMOUNT__=${KEEP_BLOCKS_AMOUNT:=8}
-__NGINX_PROTOCOL__=${NGINX_PROTOCOL:=http}
-__NGINX_HOSTNAME__=${NGINX_HOSTNAME:=localhost}
-__NGINX_PORT__=${NGINX_PORT:=8999}
-__BLOCK_WEIGHT_UNITS__=${BLOCK_WEIGHT_UNITS:=4000000}
-__MEMPOOL_BLOCKS_AMOUNT__=${MEMPOOL_BLOCKS_AMOUNT:=8}
-__BASE_MODULE__=${BASE_MODULE:=mempool}
-__ROOT_NETWORK__=${ROOT_NETWORK:=}
-__MEMPOOL_WEBSITE_URL__=${MEMPOOL_WEBSITE_URL:=https://mempool.space}
-__LIQUID_WEBSITE_URL__=${LIQUID_WEBSITE_URL:=https://liquid.network}
-__MINING_DASHBOARD__=${MINING_DASHBOARD:=true}
-__LIGHTNING__=${LIGHTNING:=false}
-__AUDIT__=${AUDIT:=false}
-__MAINNET_BLOCK_AUDIT_START_HEIGHT__=${MAINNET_BLOCK_AUDIT_START_HEIGHT:=0}
-__TESTNET_BLOCK_AUDIT_START_HEIGHT__=${TESTNET_BLOCK_AUDIT_START_HEIGHT:=0}
-__SIGNET_BLOCK_AUDIT_START_HEIGHT__=${SIGNET_BLOCK_AUDIT_START_HEIGHT:=0}
-__ACCELERATOR__=${ACCELERATOR:=false}
-__ACCELERATOR_BUTTON__=${ACCELERATOR_BUTTON:=true}
-__SERVICES_API__=${SERVICES_API:=https://mempool.space/api/v1/services}
-__PUBLIC_ACCELERATIONS__=${PUBLIC_ACCELERATIONS:=false}
-__HISTORICAL_PRICE__=${HISTORICAL_PRICE:=true}
-__ADDITIONAL_CURRENCIES__=${ADDITIONAL_CURRENCIES:=false}
-
-# Export as environment variables to be used by envsubst
-export __MAINNET_ENABLED__
-export __TESTNET_ENABLED__
-export __TESTNET4_ENABLED__
-export __SIGNET_ENABLED__
-export __LIQUID_ENABLED__
-export __LIQUID_TESTNET_ENABLED__
-export __ITEMS_PER_PAGE__
-export __KEEP_BLOCKS_AMOUNT__
-export __NGINX_PROTOCOL__
-export __NGINX_HOSTNAME__
-export __NGINX_PORT__
-export __BLOCK_WEIGHT_UNITS__
-export __MEMPOOL_BLOCKS_AMOUNT__
-export __BASE_MODULE__
-export __ROOT_NETWORK__
-export __MEMPOOL_WEBSITE_URL__
-export __LIQUID_WEBSITE_URL__
-export __MINING_DASHBOARD__
-export __LIGHTNING__
-export __AUDIT__
-export __MAINNET_BLOCK_AUDIT_START_HEIGHT__
-export __TESTNET_BLOCK_AUDIT_START_HEIGHT__
-export __SIGNET_BLOCK_AUDIT_START_HEIGHT__
-export __ACCELERATOR__
-export __ACCELERATOR_BUTTON__
-export __SERVICES_API__
-export __PUBLIC_ACCELERATIONS__
-export __HISTORICAL_PRICE__
-export __ADDITIONAL_CURRENCIES__
-
-folder=$(find /var/www/mempool -name "config.js" | xargs dirname)
-echo ${folder}
-envsubst < ${folder}/config.template.js > ${folder}/config.js
-
-exec "$@"
+set -eu
+# Preserve the pinned upstream entrypoint (including new runtime options).
+# Apply our DNS repair only after it has rendered the nginx configuration.
+exec /patch/upstream-entrypoint.sh /patch/start-nginx.sh "$@"

docker/mempool-frontend/repair-nginx.sh

@@ -0,0 +1,56 @@
+#!/bin/sh
+# Resolve the backend again after container IP changes. Run after upstream
+# placeholder substitution, so the repair also works on an existing container.
+set -eu
+conf=${1:-/etc/nginx/conf.d/nginx-mempool.conf}
+resolv=${2:-/etc/resolv.conf}
+backend=${BACKEND_MAINNET_HTTP_HOST:-127.0.0.1}
+port=${BACKEND_MAINNET_HTTP_PORT:-8999}
+resolver=$(awk '/^nameserver/ { print $2; exit }' "$resolv")
+[ -n "$resolver" ] || { echo 'No DNS resolver configured' >&2; exit 1; }
+case "$resolver" in *:*) resolver="[$resolver]" ;; esac
+case "$backend" in *[!a-zA-Z0-9._-]*|'') echo 'Invalid backend hostname' >&2; exit 1 ;; esac
+case "$port" in *[!0-9]*|'') echo 'Invalid backend port' >&2; exit 1 ;; esac
+
+tmp=$(mktemp "${conf}.archy.XXXXXX")
+trap 'rm -f "$tmp"' EXIT HUP INT TERM
+awk -v backend="$backend" -v port="$port" -v resolver="$resolver" '
+  BEGIN {
+    base = "http://" backend ":" port
+    print "# Archipelago: refresh backend DNS after container replacement."
+    print "resolver " resolver " valid=5s ipv6=off; # archy-dns"
+    print "resolver_timeout 3s; # archy-dns"
+  }
+  /# Archipelago: refresh backend DNS/ || /# archy-dns/ { next }
+  /^[[:space:]]*location[[:space:]]/ { location = $2 }
+  /^[[:space:]]*proxy_pass[[:space:]]/ && index($2, base) == 1 {
+    target = $2
+    sub(/;$/, "", target)
+    path = substr(target, length(base) + 1)
+    if (location != "/api/v1/ws" && location != "/ws" && location != "/api/v1" && location != "/api/") {
+      print "Unexpected backend location: " location > "/dev/stderr"
+      failed = 1; exit 1
+    }
+    if (path != "/" && path != "/api/v1" && path != "/api/v1/") {
+      print "Unexpected backend URI mapping" > "/dev/stderr"
+      failed = 1; exit 1
+    }
+    # Explicitly preserve prefix substitution and query arguments. A variable
+    # proxy_pass without a URI forwards the rewritten URI and original args.
+    print "\t\tset $mp_backend " backend ";"
+    if (path != location)
+      print "\t\trewrite ^" location "(.*)$ " path "$1 break;"
+    print "\t\tproxy_pass http://$mp_backend:" port ";"
+    count++
+    next
+  }
+  /proxy_pass http:\/\/\$mp_backend:/ { count++ }
+  { print }
+  END {
+    if (failed || count != 4) {
+      print "Expected four backend proxies; refusing an incomplete DNS repair" > "/dev/stderr"
+      exit 1
+    }
+  }
+' "$conf" > "$tmp"
+cat "$tmp" > "$conf"

docker/mempool-frontend/start-nginx.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -eu
+/patch/repair-nginx.sh
+nginx -t
+exec "$@"

docs/TODO.md

@@ -3,6 +3,14 @@
 Working backlog of forward-looking items not yet scoped into a dedicated plan
 doc. See [`ROADMAP.md`](ROADMAP.md) for the curated, public-facing direction.
 
+## Blocking incident — before unrelated work
+
+- **OPEN: Framework LND startup / missing Receive address / false zero balance.**
+  User requires investigation and a verified fix on the actual node before later
+  unrelated work. Access is pending; a manual LND restart is only a workaround.
+  See [incident evidence and closure criteria](incident-framework-lnd-startup.md)
+  and the repository `AGENTS.md` session-start instructions.
+
 ## Dev & build process (priority)
 
 - Formalize the contributor workflow: releases, CI, maintainers, automated

docs/incident-2026-09-15-minibits-already-redeemed.md

@@ -0,0 +1,119 @@
+# Incident — 2026-09-15: Minibits Cashu claim stuck retrying an already-redeemed token
+
+## Report
+
+User: "The cashu server is unable to get it's tokens from nostr on
+[affected node]" — clarified as the Cashu **client wallet**
+(Minibits `@minibits.cash` Lightning-address receive flow), not a mint
+server. UI showed: *"a payment arrived but couldn't be redeemed yet (1)"*.
+
+## Root cause
+
+`wallet::minibits::claim_and_redeem` (`core/archipelago/src/wallet/minibits.rs`)
+polls Nostr relays for NIP-04-encrypted Cashu tokens sent to the node's
+`@minibits.cash` address, decrypts them, and redeems them at the mint. A
+token that fails to redeem is kept in `MinibitsState.pending_claims` and
+retried on the next poll — by design, so a *transient* failure (mint briefly
+down, decrypt hiccup) never drops real money.
+
+But one queued claim had already been redeemed (mint error **11001 "Token
+Already Spent"** — most likely double-delivered by the relay, or redeemed
+by an earlier run before a crash lost track of it). That's a *terminal*
+condition, not a transient one: the code didn't distinguish the two, so it
+retried the same dead claim every ~6 seconds forever:
+
+```
+WARN archipelago::wallet::ecash: Failed to swap proofs from mint https://mint.minibits.cash/Bitcoin:
+     This ecash has already been redeemed — it can't be claimed twice.: {"code":11001,"detail":"Token Already Spent"}
+WARN archipelago::wallet::minibits: Minibits claim decrypted but failed to redeem (...); will retry next poll
+```
+
+Confirmed via `sudo journalctl -u archipelago.service` on the affected node,
+and via `/var/lib/archipelago/wallet/minibits.json`, which had exactly one
+`pending_claims` entry. Each poll also unconditionally queried all three
+`CLAIM_RELAY_URLS` (`relay.minibits.cash`, `relay.damus.io`, `nos.lol`)
+instead of the primary relay only, adding needless churn and leaking the
+wallet's Nostr pubkey to two relays it didn't need to touch — `relay.damus.io`
+was additionally failing NIP-42 auth / 503ing on every poll.
+
+**No funds were at risk** — an already-redeemed token has zero remaining
+value. The only symptom was a permanently stuck "couldn't be redeemed yet"
+banner and wasted relay connections.
+
+### Why this had already been "fixed" once and came back
+
+This exact bug (terminal-11001 handling + relay-query reduction) was fixed
+on 2026-09-09 on branch `feat/minibits-lnurl-receive` (commits `4e410d7`,
+`489995c`) and pushed to `origin`. **That branch was never merged into
+`main`.** `main` carries its own, independently-diverged rewrite of
+`minibits.rs` that never got those two hardening fixes. The affected node
+OTA'd to `1.8.16-alpha` (built from `main`) earlier on 2026-09-15, so the bug
+resurfaced on the first replayed/double-delivered claim after that update.
+
+## Fix
+
+Two parts:
+
+### 1. Immediate unstick (affected node, operational, no code change)
+
+- Backed up `/var/lib/archipelago/wallet/minibits.json`.
+- Stopped `archipelago.service`, emptied `pending_claims` (`[]`) in the
+  state file, restarted the service.
+- Verified via `journalctl` that polling resumed cleanly with no further
+  "already been redeemed" warnings.
+
+### 2. Code fix, ported into `main`
+
+- **`core/archipelago/src/wallet/mint_client.rs`**: exposed the existing
+  NUT error-code-11001 translation as a public constant,
+  `ALREADY_REDEEMED_MSG`, and a typed `AlreadyRedeemed` condition identified
+  only by the structured mint error code. Remote text cannot impersonate it.
+- **`core/archipelago/src/wallet/minibits.rs`**:
+  - Added `is_already_redeemed(&anyhow::Error) -> bool`, checking the error
+    chain for the typed `AlreadyRedeemed` condition. The ecash receive path
+    preserves it only when all failed mint entries report already-spent proofs;
+    mixed terminal/transient failures remain retryable.
+  - In the claim redeem loop, a redeem failure matching
+    `is_already_redeemed` is now dropped (logged at `info!`, not retried)
+    instead of being pushed back onto `pending_claims`. Every other failure
+    still retries next poll, unchanged.
+  - `fetch_relay_dms` now connects to `RELAY_URL` (the Minibits relay)
+    alone first via `try_connect_relay`, and only adds the two public
+    fallback relays (`relay.damus.io`, `nos.lol`) if that primary relay is
+    unreachable. Also paginates the DM fetch (200/page, capped at 5 pages)
+    backward with an inclusive `until` boundary. The cursor persists across
+    polls when capped or interrupted, independently of the forward watermark.
+    A full same-second boundary is fetched with a larger limit rather than
+    skipped, so multiple payments sharing a timestamp remain reachable.
+
+Deliberately **not** ported from the unmerged branch: its `STATE_LOCK`
+skip-if-busy guard and per-claim attempt-count backstop. `main`'s existing
+`MINIBITS_STATE_LOCK` already fully serializes claim polls (blocks rather
+than skips — a different but equally valid way to close the same race), and
+an attempt-count backstop would have required reshaping the `PendingClaim`
+enum for marginal extra protection beyond what the 11001 fix already covers.
+
+## Verification
+
+- `cargo build -p archipelago` — clean, no new warnings.
+- `cargo test -p archipelago --bin archipelago wallet::minibits` — existing
+  suite still green (see PR/commit for the run).
+- Live on the affected node: claim poll loop confirmed quiet post-unstick
+  (only `relay.minibits.cash` connects logged, no redeem-failure warnings).
+
+## Lesson (recorded in memory)
+
+A fix that lives only on an unmerged feature branch is not a fix that's
+actually deployed. Before trusting a memory or changelog claim that
+something "shipped," check which branch the running/released build was
+built from (`git log <branch>..main` / `main..<branch>`) rather than
+assuming a pushed branch was merged.
+
+## Pre-merge review regressions
+
+- A 450-event newest-first backlog is completely fetched.
+- 250 distinct payments sharing one timestamp are preserved.
+- A 1,300-event backlog resumes after the five-page cap and a state reload.
+- An interrupted relay fetch retains its unfinished cursor.
+- Only structured error 11001 is terminal, including when errors are wrapped;
+  remote free text and mixed mint failures cannot discard a retryable claim.

docs/incident-framework-lnd-startup.md

@@ -0,0 +1,378 @@
+# Framework: LND startup, missing Receive address, false zero balance
+
+**Status: OPEN — Framework startup and Cashu address verified live; source integration and final dashboard balance confirmation remain.**
+
+Reported: 2026-09-15. Source inspected: main at `3b9b74da` (v1.8.17-alpha publication).
+The Framework's installed version and exact incident time have not been verified.
+
+## Mandatory priority across sessions
+
+The user explicitly requested that this be investigated and fixed on the node
+before resuming unrelated work in later sessions. `AGENTS.md` in the repository
+and `/home/archipelago/.codex/AGENTS.md` carry this session-start priority.
+Only live verification below, or an explicit user change of priority, clears it.
+
+## Reported observations
+
+- Framework stopped showing its Lightning address in Receive.
+- After a restart, LND did not initialize and the UI displayed a balance of zero.
+- Manually restarting LND restored operation.
+- Node access will be supplied later. No Framework connection, restart, wallet
+  operation, or deployment was performed during this offline investigation.
+- Still clarify whether the restart was a full reboot or management-service
+  restart, and which Receive item vanished: a Lightning invoice, an on-chain
+  address, or the Cashu tab's `@minibits.cash` address.
+
+A successful manual restart is a workaround, not a root cause or durable fix.
+The zero display does not establish that any funds were lost. Its relation to
+v1.8.17-alpha is unknown; do not infer a release regression from timing alone.
+
+## Confirmed source findings
+
+### 1. LND errors can be presented as successful zero balances
+
+`core/archipelago/src/api/rpc/lnd/info.rs`, `handle_lnd_getinfo`:
+
+- `/v1/getinfo` is decoded without checking HTTP success. Its response fields are
+  optional, so an error object such as `{"code":14,"message":"wallet not ready"}`
+  can deserialize with every expected field absent instead of rejecting the call.
+- Channel and blockchain balance requests suppress connection/JSON failures and
+  substitute responses with absent balances. HTTP status is not checked here either.
+- Missing or unparsable balances become `0` through `unwrap_or(0)`.
+- `neode-ui/src/views/Home.vue`, `loadWeb5Status`, treats this RPC response as
+  success, sets the wallet connected flag, overwrites prior balances, and can
+  persist the false zero in the wallet snapshot. Its existing failure handling
+  preserves prior balances only when the RPC actually rejects.
+
+This is a confirmed code defect and a plausible explanation for the reported
+display. It is not proof of the Framework's failure sequence.
+
+Required fix: reject unsuccessful/incomplete LND balance responses or model
+availability explicitly end to end. Never translate unavailable data into a
+verified zero. Preserve known balances with a clear unavailable/stale indication;
+show an unknown state when no valid balance is known. Genuine successful zeros
+must still render as zero. Cover outage, partial failure, cold load, and recovery.
+
+### 2. Startup readiness and wallet unlock need live evidence
+
+- `main.rs` runs crash/container boot recovery before starting the reconciler.
+- `crash_recovery.rs` can start existing containers directly.
+- `container/prod_orchestrator.rs` runs LND post-start hooks on explicit restart
+  and on normal reconciliation of already-running containers. Therefore it is
+  incorrect to conclude that running containers categorically skip unlock.
+- `container/lnd.rs::ensure_wallet_initialized` checks wallet existence and
+  `/v1/getinfo`, then attempts unlock. Its unlock wait budget is approximately ten
+  minutes; per-request timeouts can extend elapsed time. Historical comments
+  describe slow database startup and restart loops, but that is not Framework evidence.
+- `health_monitor.rs` models LND's Bitcoin dependency. Container-running state
+  alone is not proof of wallet readiness, Bitcoin connectivity, or invoice readiness.
+
+Investigate boot ordering, Bitcoin readiness, listener/port mapping, wallet unlock,
+mount availability, stopped markers, restart counters, and actual reconcile logs.
+
+### 3. Destructive automatic recovery exists; exclude it from diagnosis
+
+`container/lnd.rs::ensure_wallet_initialized` calls
+`recreate_wallet_destructively` when all candidate passwords are rejected. That
+function can delete the LND chain and graph data directories. Its comment assumes
+alpha wallets hold no real funds; that assumption must not guide this investigation.
+
+No evidence establishes that it ran on Framework. Preserve the original wallet
+and channels; rejected passwords must lead to a recoverable error, not automatic
+wallet deletion. Review and disable this destructive fallback before using a
+modified initialization path as a repair. The existing
+`unlock_existing_wallet_no_wipe` demonstrates the non-destructive error behavior.
+
+### 4. The missing address must be identified precisely
+
+`ReceiveBitcoinModal.vue` generates Lightning invoices using `lnd.createinvoice`
+after a readiness check, and Bitcoin addresses using `lnd.newaddress`. Its Cashu
+Lightning address uses `wallet.ecash-lnaddress` and the Minibits service separately.
+Do not assume the Minibits address disappears because LND is down. Trace the actual
+tab and response once the user clarifies and the node can be inspected.
+
+## Next session: live investigation order
+
+1. Request Framework access and verify node identity without publishing its hostname,
+   address, credentials, or wallet identifiers. Do not substitute the development box.
+2. Record installed backend/image versions, boot and incident timestamps, and exact
+   restart/action sequence. Capture current and previous-boot management/LND logs
+   before another restart can obscure evidence. Keep raw logs private and redact
+   secrets, invoices, wallet identifiers, and personally identifying data in summaries.
+3. Read container/service state, restart counters, mounts, stopped markers, listener
+   mappings, Bitcoin readiness, LND wallet state, and authenticated API results.
+   Never dump container environments, macaroons, passwords, seeds, or wallet databases.
+4. Compare HTTP status and data from LND getinfo/balance endpoints with the RPC and
+   visible Receive/balance state. Distinguish unavailable data, locked wallet,
+   syncing wallet, and genuine zero. Preserve last-known balance evidence privately.
+5. Establish whether the manual restart ran a missing/failed hook, waited out a
+   dependency, refreshed networking/credentials, or masked another failure.
+6. Implement the evidenced startup repair and unavailable-balance handling with
+   regressions. Preserve wallet/channel state and arrange recovery access before
+   deploying or deliberately rebooting the node.
+
+## Acceptance criteria — all required to close
+
+- [x] Root cause of Framework startup failure supported by node evidence.
+- [x] Fix implemented and focused regression tests pass.
+- [ ] Failed, locked, delayed, and partial LND responses never masquerade as a
+      fresh zero balance; genuine zero remains correct.
+- [x] Existing wallet identity and channel state preserved through the repair.
+- [x] Framework starts LND and reaches usable wallet readiness after a controlled
+      full reboot, without manually restarting LND.
+- [ ] The originally affected Receive flow works after boot and after recovery;
+      outages show an actionable state and recover without requiring a page reload.
+- [ ] Display confirmation pending; authenticated LND balances match pre-reboot values.
+- [x] LND logs show no restart loop, repeated unlock failure, or wallet-recreation path.
+- [ ] Evidence, tested versions, deployment, and limitations recorded here; user
+      informed of live results. Only then set status RESOLVED and clear the blockers.
+
+## Work completed so far
+
+2026-09-15: source investigation and persistent session-start instructions only.
+No code fix, release, node deployment, or live reproduction for this incident yet.
+
+## Live evidence captured 2026-09-15
+
+Access was provided during the same session. Read-only inspection confirmed:
+
+- Framework runs `1.8.17-alpha-dev`; the current full boot began at 18:40:09 UTC.
+- LND opened its databases in 6.7 seconds and requested its wallet password at
+  18:40:20. It then rejected GetInfo/ChannelBalance/WalletBalance as wallet locked.
+- The management service's first sequential reconcile pass was occupied by
+  unrelated image recovery, including a missing voice image from 18:40:24 and
+  later a missing Core Lightning image. Manifests are iterated from a HashMap;
+  wallet readiness has no initial priority. Boot recovery itself completed at
+  18:40:18; the first full app-reconcile report appeared at 18:44:34.
+- The user's manual LND restart was recorded at 18:42:33. The replacement LND
+  process started at 18:42:40, requested its password at 18:43:05, and unlocked
+  at 18:43:07 through the explicit restart hook. This supports delayed unlock
+  behind unrelated recovery, rather than a missing wallet or bad password.
+- At inspection, `/v1/state` reports SERVER_ACTIVE; getinfo reports chain and
+  graph sync and two active channels. Both authenticated balance endpoints
+  report nonzero balances. No wallet-recreation event was found in captured logs.
+- The Minibits RPC separately fails with “The ecash wallet has no seed yet”.
+  `wallet/cashu_seed.json` and `wallet/minibits.json` are absent. The existing
+  ecash wallet is present with proofs and an August modification timestamp.
+  Do not overwrite it or generate an unrelated recovery identity. Still identify
+  which Receive item the user meant before declaring this part repaired.
+
+Private raw evidence: `/home/archipelago/.local/state/archy-incidents/framework-lnd-20260915/`.
+Files have mode 0600 and the directory 0700. Do not commit or publish raw logs.
+
+Candidate changes on `investigate/framework-lnd-startup`:
+
+- Run Bitcoin and LND reconciliation before unrelated image pulls/builds.
+- Reject failed/incomplete LND balance responses instead of manufacturing zeros.
+- Preserve known Home balances on invalid responses, visibly label unavailable
+  balances, and clear the warning after a successful refresh.
+- Remove automatic destructive wallet recreation; failed unlock preserves data.
+- Add backend outage/zero/ordering regressions and UI failure/recovery coverage.
+
+These changes are not yet deployed or verified through a Framework reboot.
+
+### Candidate validation and staging
+
+Source fix commit: `4237fb5e` on `investigate/framework-lnd-startup`.
+
+- 44 focused backend tests passed (including LND errors, genuine zero, startup ordering).
+- 58 additional reconciliation/update tests passed.
+- 12 Home UI tests passed, including outage/partial response/cold-load/recovery cases.
+- Rust formatting, frontend type checking and production build passed.
+- Optimized backend build passed in 8m02s.
+- Both candidate artifacts were copied to Framework and SHA-256 matched locally.
+- Private on-node baseline and static channel backup are under
+  `/var/lib/archipelago/support/framework-lnd-20260915/`, along with the previous
+  backend, dashboard, and `rollback.sh`. This directory is root-only.
+- Candidate staged at `/tmp/archy-framework-candidate/`; not applied yet.
+- A timing confirmation for the maintenance restart/full reboot was requested
+  because it interrupts all node services. Do not reboot while that is pending.
+- SSH works through the temporary control socket
+  `/tmp/archy-framework-connection/control`. No SSH password was saved to disk.
+- The supplied SSH password did not authenticate to the dashboard. Do not guess
+  additional passwords or alter dashboard authentication. Native LND diagnostics
+  are authenticated using its existing local macaroon without printing it.
+
+Status remains OPEN until deployment and live boot/Receive/balance verification.
+
+### Authorized deployment and full reboot — 2026-09-15
+
+The user answered “yes please” to applying the staged fix and rebooting. Timing
+approval is no longer pending. Applied the staged backend and dashboard after
+rechecking both checksums and rollback copies. There were no pending channel
+HTLCs at reboot. No wallet data, secrets, or recovery identities were replaced.
+
+Live results:
+
+- A different boot ID confirms a full reboot occurred.
+- Running backend on disk matches candidate SHA-256
+  `5a354f76ebe619561eef0d318e4f41f177d04004682504d7434d632733f8e298`.
+- Management service started around 19:23:57 UTC; LND asked for its wallet
+  password at 19:24:10 and logged automatic unlock at 19:24:18. No manual LND
+  restart or interactive unlock was used after this reboot.
+- LND reports SERVER_ACTIVE and chain sync. Its identity and channel-point set
+  are identical to the private pre-reboot baseline; both channels are active.
+- On-chain and Lightning balances exactly equal the pre-reboot values.
+- LND container and systemd restart counts are zero after recovery.
+- Public HTTP checks on the node returned 200 for the dashboard index and new
+  Home bundle; their bytes match the installed candidate, including the new
+  unavailable-balance notice.
+- Captured post-reboot management and LND journals in the private local evidence
+  directory. Detailed before/after identity, channel, and balance records remain
+  in the root-only support directory on Framework.
+
+The user was asked to refresh the dashboard and confirm the originally missing
+Receive item and displayed balances. Keep OPEN until that reply is assessed;
+Minibits seed absence was a separate finding and must not be mistaken for an
+LND startup failure. Candidate is a direct node deployment, not a newly signed
+fleet release. The source branch must be integrated before a subsequent release
+can preserve this fix across the fleet.
+
+### Cashu Receive follow-up
+
+The user confirmed that the remaining error is specifically on the Ecash tab:
+“Lightning address unavailable — you can still paste a token below.”
+
+Read-only checks confirm Framework has an encrypted node master seed, existing
+Cashu proofs, and neither `wallet/cashu_seed.json` nor `wallet/minibits.json`.
+The existing Minibits handler requires an ecash seed, but setup was available
+only through the Settings backup screen; Receive hid the actionable cause.
+
+UI fix commit: `a3b64670`.
+
+- Receive checks the non-secret seed status when registration fails.
+- Unseeded wallets get the existing password/TOTP/backup-passphrase-verified setup
+  component directly in Receive, with import/restore controls excluded from this
+  focused setup screen. Setup derives from the saved node seed when present.
+- The recovery words stay in the existing local reveal UI, are cleared on Done,
+  and are never emitted to Receive. Receive retries registration after Done.
+- Seeded wallets with service outages get Retry, without offering a new identity.
+- Ten focused Receive/backup tests and the production UI build passed.
+- Deployed the dashboard change without restarting services; live HTTP index and
+  setup bundle returned 200 and byte-matched the candidate.
+- Backed up original Cashu proofs to the root-only support directory as
+  `ecash-before-address-setup.json`. No seed or proof mutation was performed by
+  the assistant. Prior LND-fixed dashboard is also backed up there.
+
+The user was asked to refresh Receive → Ecash → Set up address, authenticate in
+that node UI, and click Done. Dashboard password is required to decrypt the node
+seed; the SSH password did not authenticate to the dashboard. Do not request or
+print recovery words, bypass authentication, or create an unrelated random seed.
+After completion, verify saved seed/profile presence, registration success,
+address display, and unchanged original proofs before closing the incident.
+
+### Cashu setup completed and verified — 2026-09-15
+
+The user initially reported a forgotten passphrase, then said “did it now”. No
+independent-seed fallback was implemented or used. The user completed the existing
+password-verified setup themselves; the assistant did not receive recovery words.
+
+Read-only node verification confirmed:
+
+- `wallet/cashu_seed.json` exists, is nonempty, and records source `node-seed`.
+- `wallet/minibits.json` exists with a `@minibits.cash` address and no pending claims.
+- The original ecash wallet file is byte-for-byte unchanged from the protected
+  pre-setup copy; every original proof is preserved.
+- The registered address's public LNURL-pay metadata returns HTTP 200, tag
+  `payRequest`, an HTTPS callback, and a valid amount range. No invoice was paid
+  and no funded payment test was performed.
+
+LND automatic startup and native balances were already verified after the full
+reboot. Cashu setup and address registration are now also verified on Framework.
+Do not ask for the forgotten passphrase again or propose a replacement Cashu seed.
+
+Remaining: integrate the tested source branch before the next fleet release;
+record final human confirmation of the rendered dashboard balance (native balances
+match exactly, and UI failure/recovery regressions pass). Keep this follow-up
+visible across sessions; do not rebuild/reboot/reinitialize a working wallet just
+to repeat already completed checks.
+
+### Backup copy and layout — 2026-09-15
+
+At the user's request, shortened the ecash backup explanations and stacked each
+card section's text and full-width action vertically. Kept the distinction
+between node-derived and separate phrases, and the warning that a newly created
+phrase covers future coins rather than existing legacy coins.
+
+All 10 Receive/backup tests and the production UI build pass. Deployed the UI to
+Framework without a restart; served index and backup-component bundle match the
+build byte-for-byte. The prior UI is saved as `web-ui-before-backup-copy` in the
+protected incident directory. Source integration and final rendered dashboard
+balance confirmation remain pending as above.
+
+### LNURL comment-length report — 2026-09-15
+
+User reports a maximum-comment-length error in some sending wallets. Live
+Framework address metadata advertises integer `commentAllowed: 100`. The QR
+contains the address only; Archy's Receive UI does not add a comment. The
+Minibits-hosted callback returned invoices for omitted/empty comments, 100 ASCII
+characters, 101 ASCII characters, and 100 accented characters. These were unpaid
+invoice requests at the advertised minimum amount; no funds were sent.
+
+The callback did not reproduce the error, including beyond its advertised limit.
+Sending-wallet validation against the advertised 100-character limit is therefore
+a hypothesis, not a confirmed root cause. Asked which wallets fail and whether
+an empty comment also fails. Need that result before selecting a code fix.
+The service controls the advertised limit; changing local Receive text or QR
+cannot raise it for other wallets.
+
+### Primal Spark: automatic recipient note exceeds the address limit
+
+User clarified that no comment was entered and the sender is Primal Spark.
+Checked Framework's management journal over the preceding 20 minutes: no
+comment-length errors, service active, and zero pending Minibits claims. Recent
+claim polling connected to and disconnected from the relay normally. Historical
+seed-authentication failures preceded the successful setup already documented.
+
+The live address's Minibits `text/plain` description is **101 ASCII characters**,
+while `commentAllowed` is **100**. Description template (address redacted):
+`Pay to [ADDRESS] with Lightning. Receiver will receive ecash into Minibits Wallet.`
+
+Primal Android source at `36939db97213e7f8eeefaa4adaf125d839fc662e`:
+- `WalletTextParserImpl.handleLnUrlText` assigns the parsed description to
+  `DraftTx.noteRecipient`, including for Lightning-address input.
+- `TransactionEditor` initializes its editable recipient note from that value.
+- `SparkWalletServiceImpl` passes it untrimmed to `PrepareLnurlPayRequest.comment`.
+- Breez Spark source at `8bb38ec292a590907360c4e7f2a4134b8f09de9e`,
+  `common/src/lnurl/pay.rs::validate_user_input`, rejects a comment exceeding the
+  limit with the exact reported error before requesting the callback.
+
+This identifies a concrete compatibility failure: the address description can
+become an automatic over-limit comment without the sender typing anything.
+The user confirmed that explicitly clearing the prefilled recipient note made
+the payment work, and supplied the same description observed in live metadata.
+This confirms the automatic-comment compatibility failure. The installed Primal
+platform/version was not captured. Node logs alone cannot show sender-side
+validation or requests to the external Minibits callback.
+
+Durable upstream correction: Primal should keep receiver metadata separate from
+the sender's comment and enforce the limit on actual user comments. Minibits can
+also shorten its description or raise its advertised comment limit. Archy does
+not serve this external LNURL metadata; do not rename an existing wallet address,
+rotate its seed, or claim that a local dashboard edit fixes this sender behavior.
+
+### Primal workaround confirmed by user
+
+The user confirmed successful payment after removing the automatic description.
+The permanent sender-side correction is to leave the recipient comment empty by
+default and retain receiver metadata only as display text. In Primal Android,
+remove the assignment of the LNURL description to the draft recipient note in
+`WalletTextParserImpl.handleLnUrlText`; also validate explicitly entered comments
+against the endpoint's limit. No upstream change has been submitted or deployed.
+Existing Framework addresses and wallet identities remain unchanged.
+
+### Can Archy shorten the current address description?
+
+Inspected Minibits' public wallet client (`src/services/minibitsService.ts`,
+`updateWalletProfile`) and `WalletProfileRecord`. The supported profile update
+fields are name, lud16, and avatar; there is no exposed LNURL description or
+comment-limit setting. Its public web repository also contains no implementation
+of the LNURL metadata endpoint or description template.
+
+For the existing `@minibits.cash` address, no supported client-side mechanism
+to shorten this text was found. Do not send guessed profile-update fields or
+rename the address to disguise the problem. A Minibits server change could use
+`Pay to [ADDRESS]`, well below the current limit. Controlling this metadata in
+Archy would instead require an Archy-hosted LNURL service/address and correct
+invoice metadata binding; rewriting the QR label or only proxying edited metadata
+is insufficient. No wallet/profile mutations were made during this investigation.

neode-ui/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "neode-ui",
-  "version": "1.8.14-alpha",
+  "version": "1.8.17-alpha",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "neode-ui",
-      "version": "1.8.14-alpha",
+      "version": "1.8.17-alpha",
       "dependencies": {
         "@scure/bip39": "^2.2.0",
         "@types/dompurify": "^3.0.5",

neode-ui/package.json

@@ -1,7 +1,7 @@
 {
   "name": "neode-ui",
   "private": true,
-  "version": "1.8.14-alpha",
+  "version": "1.8.17-alpha",
   "type": "module",
   "scripts": {
     "start": "./start-dev.sh",

neode-ui/public/catalog.json

@@ -378,13 +378,13 @@
     {
       "id": "mempool",
       "title": "Mempool Explorer",
-      "version": "3.0.0",
+      "version": "3.3.1-archy1",
       "description": "Bitcoin mempool and blockchain explorer. Real-time transaction and block visualization.",
       "icon": "/assets/img/app-icons/mempool.webp",
       "author": "Mempool",
       "category": "money",
       "tier": "core",
-      "dockerImage": "source.archipelago-foundation.org/lfg2025/mempool-frontend:v3.3.1",
+      "dockerImage": "source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1",
       "repoUrl": "https://github.com/mempool/mempool",
       "requires": [
         "bitcoin-knots",

neode-ui/src/components/EcashSeedBackup.vue

@@ -3,6 +3,9 @@ import { ref, computed, onMounted } from 'vue'
 import { rpcClient } from '@/api/rpc-client'
 import SeedRevealPanel from '@/components/SeedRevealPanel.vue'
 
+defineProps<{ setupOnly?: boolean }>()
+const emit = defineEmits<{ ready: [] }>()
+
 // Ecash (Cashu) wallet backup card — the same shape as the node recovery
 // phrase and the Lightning seed cards, deliberately: a third reveal pattern
 // would be a third thing to learn.
@@ -102,12 +105,14 @@ async function submitReveal() {
 }
 
 function closeReveal() {
+  const established = revealedWords.value.length > 0
   showRevealModal.value = false
   revealedWords.value = []
   revealPassword.value = ''
   revealCode.value = ''
   revealPassphrase.value = ''
   showRevealPassphrase.value = false
+  if (established) emit('ready')
 }
 
 async function copyRevealedWords() {
@@ -221,61 +226,54 @@ async function restoreFromPhrase() {
       Your ecash has no backup yet
     </div>
 
-    <div class="flex items-start justify-between gap-4">
+    <div class="flex flex-col gap-3">
       <div class="min-w-0">
-        <h2 class="text-xl font-semibold text-white/96 mb-1">Ecash backup phrase</h2>
+        <h2 class="text-xl font-semibold text-white/96 mb-1">{{ setupOnly ? 'Set up your Cashu Lightning address' : 'Ecash backup phrase' }}</h2>
 
-        <p v-if="status?.active && status?.source === 'node-seed'" class="text-sm text-white/60">
-          Your ecash wallet has its own 24-word phrase, derived from this node's recovery
-          phrase — so the words you already wrote down cover your ecash too. Reveal it here
-          if you want to restore your ecash into another wallet (Minibits, Nutstash,
-          <span class="font-mono">cdk-cli</span>) without handing over the node's own seed.
+        <p v-if="status?.active && status?.source === 'node-seed'" class="text-sm leading-relaxed text-white/60">
+          Your node's recovery phrase also recovers this ecash phrase. Reveal its 24 words
+          to restore in a compatible Cashu wallet without sharing your node's phrase.
         </p>
-        <p v-else-if="status?.active" class="text-sm text-white/60">
-          Your ecash wallet has its own 24-word phrase. Reveal it to write it down, or to
-          restore your ecash into another wallet (Minibits, Nutstash,
-          <span class="font-mono">cdk-cli</span>).
+        <p v-else-if="status?.active" class="text-sm leading-relaxed text-white/60">
+          Save your 24-word ecash phrase to restore this wallet here or in another
+          compatible Cashu wallet.
         </p>
-        <p v-else class="text-sm text-white/60">
-          Ecash is a bearer instrument: the coins live in a file on this node, and right now
-          nothing can bring them back if that file is lost. Setting up a backup phrase fixes
-          that for every coin minted from then on.
+        <p v-else class="text-sm leading-relaxed text-white/60">
+          If this node's coin file is lost, your ecash is lost. Set up a phrase to recover
+          future coins; existing coins aren't covered.
           <template v-if="status?.derivable_from_node_seed">
-            It's derived from this node's recovery phrase, so there's nothing new to write down.
+            Your node's recovery phrase will also recover this phrase.
           </template>
           <template v-else>
-            This node has no encrypted seed backup to derive from, so the phrase will be its
-            own — you'll need to write these words down and keep them.
+            This node has no saved seed, so write down and keep the new phrase separately.
           </template>
         </p>
 
         <p v-if="status?.source === 'independent' || status?.source === 'imported'" class="mt-2 text-xs text-orange-300/90">
-          This wallet's phrase was <strong>not</strong> derived from the node's recovery
-          phrase{{ status?.source === 'imported' ? ' — it was imported' : '' }}, so restoring
-          the node will not bring the ecash back. Only these words will.
+          {{ status?.source === 'imported' ? 'This imported phrase' : 'This phrase' }} is separate
+          from your node's backup. <strong>Only these words recover this ecash wallet.</strong>
         </p>
 
       </div>
 
       <button
         type="button"
-        class="shrink-0 glass-button rounded-lg px-4 py-2 text-sm font-medium"
+        class="w-full glass-button rounded-lg px-4 py-2 text-sm font-medium"
         :class="!status?.active ? 'bg-orange-500/20 border-orange-400/30' : ''"
         @click="openReveal"
-      >{{ status?.active ? 'Reveal' : 'Set up backup' }}</button>
+      >{{ status?.active ? 'Reveal' : (setupOnly ? 'Set up address' : 'Set up backup') }}</button>
     </div>
 
-    <div v-if="status?.active" class="mt-4 pt-4 border-t border-white/10">
-      <div class="flex items-start justify-between gap-4">
-        <p class="text-sm text-white/60 min-w-0">
+    <div v-if="status?.active && !setupOnly" class="mt-4 pt-4 border-t border-white/10">
+      <div class="flex flex-col gap-3">
+        <p class="text-sm leading-relaxed text-white/60 min-w-0">
           <span class="text-white/80 font-medium">Restore from this phrase.</span>
-          Asks your mint which coins it has signed for these words and puts back any that
-          are still unspent. Safe to run at any time — it never duplicates coins you already
-          hold.
+          Recover unspent coins from your mint. Safe to repeat; coins you already hold
+          won't be duplicated.
         </p>
         <button
           type="button"
-          class="shrink-0 glass-button rounded-lg px-4 py-2 text-sm font-medium disabled:opacity-50"
+          class="w-full glass-button rounded-lg px-4 py-2 text-sm font-medium disabled:opacity-50"
           :disabled="restoring"
           @click="restoreFromPhrase"
         >{{ restoring ? 'Scanning…' : 'Restore' }}</button>
@@ -284,16 +282,16 @@ async function restoreFromPhrase() {
       <p v-if="restoreError" role="alert" class="mt-3 text-xs alert-error px-3 py-2 rounded-lg">{{ restoreError }}</p>
     </div>
 
-    <div class="mt-4 pt-4 border-t border-white/10">
-      <div class="flex items-start justify-between gap-4">
-        <p class="text-sm text-white/60 min-w-0">
+    <div v-if="!setupOnly" class="mt-4 pt-4 border-t border-white/10">
+      <div class="flex flex-col gap-3">
+        <p class="text-sm leading-relaxed text-white/60 min-w-0">
           <span class="text-white/80 font-medium">Use a phrase from another wallet.</span>
-          Point this wallet at a phrase you already have — from Minibits, Nutstash or
-          <span class="font-mono">cdk-cli</span> — so its coins can be restored here.
+          Import a phrase from Minibits, Nutstash or <span class="font-mono">cdk-cli</span>
+          to restore its coins here.
         </p>
         <button
           type="button"
-          class="shrink-0 glass-button rounded-lg px-4 py-2 text-sm font-medium"
+          class="w-full glass-button rounded-lg px-4 py-2 text-sm font-medium"
           @click="openImport"
         >Import</button>
       </div>
@@ -319,7 +317,7 @@ async function restoreFromPhrase() {
         </template>
 
         <template v-else>
-          <p class="text-sm text-white/60 mb-4">
+          <p class="text-sm leading-relaxed text-white/60 mb-4">
             Paste the 24-word phrase from the other wallet. The coins already in this wallet
             stay spendable either way.
           </p>
@@ -376,9 +374,8 @@ async function restoreFromPhrase() {
         </h3>
 
         <template v-if="revealedWords.length === 0">
-          <p class="text-sm text-white/60 mb-4">
-            Confirm your credentials to
-            {{ status?.active ? 'display the 24-word ecash phrase' : 'derive and display your ecash backup phrase' }}.
+          <p class="text-sm leading-relaxed text-white/60 mb-4">
+            Confirm your credentials to {{ status?.active ? 'reveal' : 'set up' }} your ecash phrase.
           </p>
           <form @submit.prevent="submitReveal" class="space-y-3">
             <div>
@@ -407,12 +404,12 @@ async function restoreFromPhrase() {
           <SeedRevealPanel :words="revealedWords" />
           <p class="text-xs text-white/40 mt-3">
             <template v-if="revealedSource === 'node-seed'">
-              Derived from this node's recovery phrase — restoring the node restores this
-              ecash wallet too. These words also restore it into any NUT-13 wallet.
+              Your node's recovery phrase recovers this ecash wallet too. Use these words
+              separately in a compatible Cashu (NUT-13) wallet.
             </template>
             <template v-else>
-              This phrase is independent of the node's recovery phrase. It is the
-              <strong>only</strong> way to restore this ecash wallet — write it down.
+              Write these words down. They are the <strong>only</strong> way to recover
+              this ecash wallet; your node's phrase won't recover it.
             </template>
           </p>
           <div class="flex gap-2 pt-4">

neode-ui/src/components/ReceiveBitcoinModal.vue

@@ -77,8 +77,13 @@
           <div v-else-if="lnAddressLoading" class="mb-4 text-center text-white/50 text-sm py-4">
             {{ t('receiveBitcoin.lnAddressLoading') }}
           </div>
+          <div v-else-if="lnAddressNeedsSetup" class="mb-3">
+            <p class="text-sm text-white/70 mb-3">Set up this wallet's recovery phrase once to enable its Lightning address.</p>
+            <EcashSeedBackup setup-only @ready="loadLnAddress" />
+          </div>
           <div v-else-if="lnAddressError" class="mb-3 text-xs text-white/40">
             {{ t('receiveBitcoin.lnAddressUnavailable') }}
+            <button type="button" class="glass-button rounded-lg px-3 py-2 ml-2" @click="loadLnAddress">Retry</button>
           </div>
 
           <div class="mb-3">
@@ -132,6 +137,7 @@ import { useI18n } from 'vue-i18n'
 import { rpcClient } from '@/api/rpc-client'
 import BaseModal from '@/components/BaseModal.vue'
 import CopyButton from '@/components/CopyButton.vue'
+import EcashSeedBackup from '@/components/EcashSeedBackup.vue'
 import PaymentSuccessPane, { type SuccessRow } from '@/components/PaymentSuccessPane.vue'
 import { explainReceiveAddressFailure } from '@/utils/bitcoinReceive'
 import { useLightningRequired } from '@/composables/useLightningRequired'
@@ -214,6 +220,7 @@ const error = ref('')
 const lnAddress = ref('')
 const lnAddressLoading = ref(false)
 const lnAddressError = ref(false)
+const lnAddressNeedsSetup = ref(false)
 // A payment the backend fetched (and so already consumed at Minibits) but
 // couldn't redeem yet — it's queued for automatic retry, not lost, but the
 // operator should see it rather than have it be a silent, unbounded wait.
@@ -230,6 +237,7 @@ async function loadLnAddress() {
   if (lnAddress.value || lnAddressLoading.value) return
   lnAddressLoading.value = true
   lnAddressError.value = false
+  lnAddressNeedsSetup.value = false
   try {
     const res = await rpcClient.call<{ address?: string }>({
       method: 'wallet.ecash-lnaddress',
@@ -245,6 +253,16 @@ async function loadLnAddress() {
     }
   } catch {
     lnAddressError.value = true
+    // A legacy wallet may hold valid proofs without having a recovery phrase.
+    // Use the existing authenticated setup flow; never silently create a new
+    // identity or send the user to an unexplained generic service error.
+    try {
+      const seedStatus = await rpcClient.call<{ active: boolean; can_activate: boolean }>({
+        method: 'wallet.ecash-seed-status',
+        timeout: 5000,
+      })
+      lnAddressNeedsSetup.value = seedStatus.active === false && seedStatus.can_activate === true
+    } catch { /* Keep the retryable service error when status is unavailable. */ }
   } finally {
     lnAddressLoading.value = false
   }

neode-ui/src/components/__tests__/EcashSeedBackup.test.ts

@@ -22,6 +22,37 @@ describe('EcashSeedBackup reveal credentials (#127)', () => {
     document.body.innerHTML = ''
   })
 
+  it('signals readiness only after authenticated setup is finished and clears the words', async () => {
+    vi.mocked(rpcClient.call).mockImplementation(async ({ method }) => {
+      if (method === 'wallet.ecash-seed-status') {
+        return { active: false, can_activate: true, derivable_from_node_seed: true, source: null } as never
+      }
+      if (method === 'wallet.ecash-seed-reveal') {
+        return { words: [...Array(23).fill('abandon'), 'art'], source: 'node-seed' } as never
+      }
+      throw new Error('unexpected request')
+    })
+    wrapper = mount(EcashSeedBackup, { props: { setupOnly: true }, attachTo: document.body })
+    await flushPromises()
+    await wrapper.get('button').trigger('click')
+    const cancel = Array.from(document.body.querySelectorAll('button')).find(b => b.textContent === 'Cancel')!
+    cancel.click()
+    await flushPromises()
+    expect(wrapper.emitted('ready')).toBeUndefined()
+    await wrapper.get('button').trigger('click')
+    const password = document.body.querySelector<HTMLInputElement>('input[autocomplete="current-password"]')!
+    password.value = 'test-password'
+    password.dispatchEvent(new Event('input', { bubbles: true }))
+    document.body.querySelector('form')!.dispatchEvent(new Event('submit', { bubbles: true, cancelable: true }))
+    await flushPromises()
+    expect(wrapper.emitted('ready')).toBeUndefined()
+    Array.from(document.body.querySelectorAll('button')).find(b => b.textContent === 'Done')!.click()
+    await flushPromises()
+    expect(wrapper.emitted('ready')).toEqual([[]])
+    expect(document.body.querySelector('[aria-labelledby="reveal-ecash-seed-title"]')).toBeNull()
+    expect(document.body.textContent).not.toContain('abandon')
+  })
+
   it('asks for a separate backup passphrase only after password decryption fails', async () => {
     vi.mocked(rpcClient.call)
       .mockResolvedValueOnce({

neode-ui/src/components/__tests__/ReceiveBitcoinModal.test.ts

@@ -1,6 +1,7 @@
 import { flushPromises, mount } from '@vue/test-utils'
 import { beforeEach, describe, expect, it, vi } from 'vitest'
 import ReceiveBitcoinModal from '../ReceiveBitcoinModal.vue'
+import EcashSeedBackup from '../EcashSeedBackup.vue'
 import { rpcClient } from '@/api/rpc-client'
 
 vi.mock('vue-router', () => ({
@@ -39,6 +40,51 @@ beforeEach(() => {
 // unmounts the dialog — but the RPC-eager tab switch is exactly the kind of
 // path a future change could regress, so it's worth pinning down.
 describe('ReceiveBitcoinModal — ecash tab click', () => {
+  it('offers authenticated setup for an unseeded wallet and retries the address after setup', async () => {
+    let active = false
+    vi.mocked(rpcClient.call).mockImplementation(async ({ method }) => {
+      if (method === 'wallet.ecash-lnaddress') {
+        if (!active) throw new Error('The ecash wallet has no seed yet')
+        return { address: 'someone@minibits.cash' } as never
+      }
+      if (method === 'wallet.ecash-seed-status') {
+        return { active, can_activate: true, derivable_from_node_seed: true, source: null } as never
+      }
+      return {} as never
+    })
+    const wrapper = mount(ReceiveBitcoinModal, { props: { show: true }, attachTo: document.body })
+    const tab = Array.from(document.body.querySelectorAll('button')).find(b => b.textContent?.toLowerCase().includes('ecash'))!
+    tab.click()
+    await flushPromises()
+    expect(document.body.textContent).toContain('Set up your Cashu Lightning address')
+    expect(document.body.textContent).not.toContain('receiveBitcoin.lnAddressUnavailable')
+    expect(vi.mocked(rpcClient.call).mock.calls.some(([r]) => r.method === 'wallet.ecash-seed-reveal')).toBe(false)
+    active = true
+    wrapper.findComponent(EcashSeedBackup).vm.$emit('ready')
+    await flushPromises()
+    expect(document.body.textContent).toContain('someone@minibits.cash')
+    expect(wrapper.emitted('close')).toBeFalsy()
+    wrapper.unmount()
+  })
+
+  it('keeps a seeded wallet on the retry path during a service outage', async () => {
+    vi.mocked(rpcClient.call).mockImplementation(async ({ method }) => {
+      if (method === 'wallet.ecash-seed-status') return { active: true, can_activate: true } as never
+      throw new Error('service unavailable')
+    })
+    const wrapper = mount(ReceiveBitcoinModal, { props: { show: true }, attachTo: document.body })
+    Array.from(document.body.querySelectorAll('button')).find(b => b.textContent?.toLowerCase().includes('ecash'))!.click()
+    await flushPromises()
+    expect(wrapper.findComponent(EcashSeedBackup).exists()).toBe(false)
+    expect(document.body.textContent).toContain('receiveBitcoin.lnAddressUnavailable')
+    const retry = Array.from(document.body.querySelectorAll('button')).find(b => b.textContent === 'Retry')!
+    expect(retry).toBeTruthy()
+    retry.click()
+    await flushPromises()
+    expect(vi.mocked(rpcClient.call).mock.calls.filter(([r]) => r.method === 'wallet.ecash-lnaddress')).toHaveLength(2)
+    wrapper.unmount()
+  })
+
   it('does not close/emit when the ecash tab is clicked and the RPC succeeds', async () => {
     vi.mocked(rpcClient.call).mockResolvedValue({ address: 'someone@minibits.cash' } as never)
 

neode-ui/src/views/Home.vue

@@ -133,6 +133,7 @@
           class="order-2 lg:order-none"
           :animate="animateCards"
           :wallet-connected="walletConnected"
+          :wallet-balance-unavailable="walletBalanceUnavailable"
           :wallet-onchain="walletOnchain"
           :wallet-lightning="walletLightning"
           :wallet-ecash="walletEcash"
@@ -685,6 +686,7 @@ async function devFaucet() { try { await rpcClient.call({ method: 'dev.faucet',
 // readout instead; a rail only becomes a number when a call actually
 // succeeds, so a real 0 is still a real 0.
 const walletConnected = ref(false)
+const walletBalanceUnavailable = ref(false)
 const walletOnchain = ref<number | null>(null)
 const walletLightning = ref<number | null>(null)
 const walletEcash = ref<number | null>(null)
@@ -775,13 +777,24 @@ async function loadWeb5Status() {
   // call, which is what makes the card feel like an app launch.
   const balances = Promise.allSettled([
     rpcClient.call<{ balance_sats: number; channel_balance_sats: number }>({ method: 'lnd.getinfo', timeout: 5000, dedup: true })
-      .then(res => { walletOnchain.value = res.balance_sats || 0; walletLightning.value = res.channel_balance_sats || 0; walletConnected.value = true; walletInfoFailures = 0 })
+      .then(res => {
+        if (!Number.isSafeInteger(res.balance_sats) || res.balance_sats < 0 ||
+            !Number.isSafeInteger(res.channel_balance_sats) || res.channel_balance_sats < 0) {
+          throw new Error('LND balance is unavailable')
+        }
+        walletOnchain.value = res.balance_sats
+        walletLightning.value = res.channel_balance_sats
+        walletConnected.value = true
+        walletBalanceUnavailable.value = false
+        walletInfoFailures = 0
+      })
       .catch(() => {
         // A single slow poll must NOT flip the card to "disconnected" and
         // hide balances the user already knows — busy nodes routinely blow
         // the 5s budget mid-payment or during IO storms (a test node user
         // report: balances vanished while a payment settled). Only call it
         // disconnected after three consecutive failures (~30s of silence).
+        walletBalanceUnavailable.value = true
         walletInfoFailures += 1
         if (walletInfoFailures >= 3) walletConnected.value = false
       }),

neode-ui/src/views/__tests__/homeTabCache.test.ts

@@ -238,6 +238,47 @@ describe('Home tab cache (Task 2): system/update/storage groups + wallet freshne
     wrapper.unmount()
   })
 
+  it.each(['failure', 'missing', 'partial'])('preserves known balances during %s and clears the warning on recovery', async (failure) => {
+    const wrapper = mountHomeHost()
+    await settle()
+    const home = wrapper.findComponent(Home)
+    const refresh = () => (home.vm as unknown as { loadWeb5Status: () => Promise<void> }).loadWeb5Status()
+    rpcCallMock.mockImplementationOnce(async () => {
+      if (failure === 'failure') throw new Error('wallet locked')
+      return failure === 'partial' ? { balance_sats: 0 } : {}
+    })
+    await refresh()
+    await settle()
+    const card = wrapper.findComponent(HomeWalletCard)
+    expect(card.props('walletOnchain')).toBe(5000)
+    expect(card.props('walletLightning')).toBe(2500)
+    expect(card.find('[data-testid="wallet-balance-unavailable"]').text()).toContain('last known')
+    const snapshot = JSON.parse(localStorage.getItem('archy-wallet-snapshot-v1')!)
+    expect(snapshot.onchain).toBe(5000)
+    expect(snapshot.lightning).toBe(2500)
+    rpcCallMock.mockImplementationOnce(async () => ({ balance_sats: 0, channel_balance_sats: 0, synced_to_chain: true }))
+    await refresh()
+    await settle()
+    expect(card.props('walletOnchain')).toBe(0)
+    expect(card.props('walletLightning')).toBe(0)
+    expect(card.find('[data-testid="wallet-balance-unavailable"]').exists()).toBe(false)
+    wrapper.unmount()
+  })
+
+  it('shows unknown rather than zero when the first LND request fails', async () => {
+    rpcCallMock.mockImplementation(async (request) => {
+      if (request.method === 'lnd.getinfo') throw new Error('wallet locked')
+      return defaultRpcCallImpl(request)
+    })
+    const wrapper = mountHomeHost()
+    await settle()
+    const card = wrapper.findComponent(HomeWalletCard)
+    expect(card.props('walletOnchain')).toBeNull()
+    expect(card.props('walletLightning')).toBeNull()
+    expect(card.find('[data-testid="wallet-balance-unavailable"]').text()).toContain('unavailable')
+    wrapper.unmount()
+  })
+
   it('no sessionStorage key exists for the wallet resource after a mount and reactivation cycle', async () => {
     const wrapper = mountHomeHost()
     await settle()

neode-ui/src/views/apps/__tests__/appsConfig.test.ts

@@ -78,12 +78,11 @@ describe('appsConfig service filtering', () => {
 
   it('shows Cuprate as one My Apps entry while hiding its daemon dependency', () => {
     const entries: Array<[string, PackageDataEntry]> = [
-      ['cuprate-ui', makePkg('cuprate-ui', 'Cuprate UI', 'money')],
-      ['cuprate', makePkg('cuprate', 'Cuprate daemon', 'money')],
+      ['cuprate', makePkg('cuprate', 'Cuprate', 'money')],
+      ['archy-cuprate-ui', makePkg('archy-cuprate-ui', 'Cuprate UI companion', 'money')],
     ]
-    ;(entries[0]![1].manifest as unknown as Record<string, unknown>).interfaces = { main: { ui: 'http://localhost:18091' } }
-    expect(filterEntriesForTab(entries, 'apps', 'all').map(([id]) => id)).toEqual(['cuprate-ui'])
-    expect(filterEntriesForTab(entries, 'services', 'all').map(([id]) => id)).toEqual(['cuprate'])
+    expect(filterEntriesForTab(entries, 'apps', 'all').map(([id]) => id)).toEqual(['cuprate'])
+    expect(filterEntriesForTab(entries, 'services', 'all').map(([id]) => id)).toEqual([])
   })
 
   it('falls back to packaged app icon when static icon token is not a path', () => {

neode-ui/src/views/apps/appsConfig.ts

@@ -20,6 +20,10 @@ export const isServiceContainer = sharedIsServiceContainer
 
 const INTERNAL_TOOLING_NAMES = new Set([
   'buildx_buildkit_default',
+  // Cuprate's dashboard is bundled as a companion of the primary cuprate
+  // package; showing the generated container as a second Services entry
+  // defeats the one-app presentation.
+  'archy-cuprate-ui',
 ])
 
 export function isInternalToolingPackage(id: string, pkg?: PackageDataEntry): boolean {
@@ -37,7 +41,7 @@ export function isServicePackage(id: string, pkg?: PackageDataEntry): boolean {
 
 // Known app -> category mappings (matches App Store categorisation)
 export const APP_CATEGORY_MAP: Record<string, string> = {
-  'bitcoin-core': 'money', 'bitcoin-knots': 'money', 'bitcoin-ui': 'money', 'cuprate-ui': 'money', 'electrumx': 'money', 'electrs': 'money',
+  'bitcoin-core': 'money', 'bitcoin-knots': 'money', 'bitcoin-ui': 'money', 'cuprate': 'money', 'cuprate-ui': 'money', 'electrumx': 'money', 'electrs': 'money',
   'lnd': 'money', 'mempool': 'money', 'mempool-web': 'money', 'btcpay-server': 'commerce',
   'fedimint': 'money', 'fedimint-gateway': 'money',
   'indeedhub': 'media', 'jellyfin': 'media', 'photoprism': 'media', 'immich': 'media',

neode-ui/src/views/apps/serviceNames.ts

@@ -14,8 +14,10 @@
 // SERVICE_NAMES set that used to live in appsConfig.ts verbatim.
 export const SERVICE_NAMES = new Set([
   'dwn', 'archy-mempool-db', 'archy-btcpay-db', 'archy-nbxplorer', 'archy-tor',
-  // Cuprate's daemon is the backend dependency of the Cuprate UI app.
-  'cuprate',
+  // Cuprate is presented as one user-facing app. Its companion container
+  // (archy-cuprate-ui) is the implementation detail; do not classify the
+  // primary package as a Service or it disappears from My Apps and loses its
+  // launch button during/after install.
   // Headless backends with no user-facing UI: the Fedimint ecash client daemon,
   // the Nostr relay, and the Meshtastic LoRa daemon (its chat UI lives in the
   // built-in Mesh tab) belong in Services, not My Apps.

neode-ui/src/views/home/HomeWalletCard.vue

@@ -54,6 +54,12 @@
           </div>
         </div>
 
+        <p v-if="walletBalanceUnavailable" data-testid="wallet-balance-unavailable" class="text-sm text-amber-200 mb-3" role="status">
+          {{ walletOnchain != null || walletLightning != null
+            ? 'Bitcoin and Lightning balances could not be refreshed. Showing last known amounts.'
+            : 'Bitcoin and Lightning balances are unavailable while the wallet starts or reconnects.' }}
+        </p>
+
         <!-- Incoming Transactions Panel -->
         <transition name="incoming-tx-slide">
           <div v-if="showIncomingTxPanel && incomingTransactions.length > 0" class="mb-4 rounded-xl overflow-hidden border border-green-500/20">
@@ -221,6 +227,7 @@ export interface WalletTransaction {
 const props = defineProps<{
   animate: boolean
   walletConnected: boolean
+  walletBalanceUnavailable?: boolean
   // `null` = not loaded yet, `0` = genuinely empty. Keeping those apart is
   // what lets the card show a pixel readout instead of claiming a figure.
   walletOnchain: number | null

neode-ui/src/views/settings/AccountInfoSection.vue

@@ -362,6 +362,44 @@ init()
               </button>
             </div>
             <div class="overflow-y-auto flex-1 min-h-0 space-y-6 pr-1">
+              <!-- v1.8.17-alpha -->
+              <div>
+                <div class="flex items-center gap-2 mb-3">
+                  <span class="text-xs font-mono px-2 py-0.5 rounded bg-orange-500/20 text-orange-300">v1.8.17-alpha</span>
+                  <span class="text-xs text-white/40">September 15, 2026</span>
+                </div>
+                <div class="space-y-3 text-sm text-white/80 pl-3 border-l border-white/10">
+                  <p>Minibits claims that every mint reports as already spent leave the retry queue, clearing repeated failure notices. Network errors and mixed mint failures remain queued for another attempt.</p>
+                  <p>Minibits polls its primary relay first and connects to public fallback relays only when the primary is unreachable, reducing unnecessary connections.</p>
+                  <p>Large payment backlogs are fetched from newest to oldest with a saved cursor, so polling can resume after interruptions or page limits. Payments sharing the same timestamp remain reachable.</p>
+                  <p>Added regression coverage for spent-claim classification, wrapped and mixed mint errors, same-second payments, and interrupted or multi-poll backlogs.</p>
+                </div>
+              </div>
+              <!-- v1.8.16-alpha -->
+              <div>
+                <div class="flex items-center gap-2 mb-3">
+                  <span class="text-xs font-mono px-2 py-0.5 rounded bg-orange-500/20 text-orange-300">v1.8.16-alpha</span>
+                  <span class="text-xs text-white/40">September 15, 2026</span>
+                </div>
+                <div class="space-y-3 text-sm text-white/80 pl-3 border-l border-white/10">
+                  <p>App updates refresh and verify the signed catalog before changing containers. A failed refresh or manifest reload cancels the update, and automatic updates wait for a successful refresh.</p>
+                  <p>Fixed repeated Mempool update offers: downstream -archyN patches now sort above their upstream release, and moving a published image between registry namespaces does not hide a genuine upgrade.</p>
+                  <p>Updates inspect installed component versions, refuse known downgrades, skip containers already at the target versions, and verify the resulting versions before reporting success.</p>
+                  <p>Added regression coverage for stale catalogs, matching versions, publisher namespace changes, stack component updates, and keeping running containers untouched when no upgrade is needed.</p>
+                </div>
+              </div>
+              <!-- v1.8.15-alpha -->
+              <div>
+                <div class="flex items-center gap-2 mb-3">
+                  <span class="text-xs font-mono px-2 py-0.5 rounded bg-orange-500/20 text-orange-300">v1.8.15-alpha</span>
+                  <span class="text-xs text-white/40">September 13, 2026</span>
+                </div>
+                <div class="space-y-3 text-sm text-white/80 pl-3 border-l border-white/10">
+                  <p>Cuprate is presented as one user-facing app in My Apps, including its UI launch button; the generated dashboard companion is hidden as an implementation detail instead of appearing under Services.</p>
+                  <p>Added regression coverage for Cuprate install and installed-state grouping.</p>
+                  <p>Release validation was rerun on the corrected tree before OTA and ISO publication.</p>
+                </div>
+              </div>
               <!-- v1.8.14-alpha -->
               <div>
                 <div class="flex items-center gap-2 mb-3">

release-manifest.json

@@ -1,29 +1,30 @@
 {
   "changelog": [
-    "**GitWorkshop installs reliably on fresh nodes.** The app is classified as a user-facing app while its install placeholder is being created, so it remains visible under My Apps instead of Services.",
-    "**Fresh GitWorkshop installs build the correct image.** The production orchestrator handles its bundled build context instead of sending the local image reference through the legacy registry-pull path.",
-    "**Curated app classification is regression-tested.** Every user-facing app remains in My Apps during installation, while headless services stay in Services."
+    "Minibits claims that every mint reports as already spent leave the retry queue, clearing repeated failure notices. Network errors and mixed mint failures remain queued for another attempt.",
+    "Minibits polls its primary relay first and connects to public fallback relays only when the primary is unreachable, reducing unnecessary connections.",
+    "Large payment backlogs are fetched from newest to oldest with a saved cursor, so polling can resume after interruptions or page limits. Payments sharing the same timestamp remain reachable.",
+    "Added regression coverage for spent-claim classification, wrapped and mixed mint errors, same-second payments, and interrupted or multi-poll backlogs."
   ],
   "components": [
     {
-      "current_version": "1.8.13-alpha",
-      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.13-alpha/archipelago",
+      "current_version": "1.8.17-alpha",
+      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.17-alpha/archipelago",
       "name": "archipelago",
-      "new_version": "1.8.13-alpha",
-      "sha256": "832c7e75b395f94f919a21d1ad7d32d367e1ef84a0e0995b4e8fe87268aa21a4",
-      "size_bytes": 64585424
+      "new_version": "1.8.17-alpha",
+      "sha256": "32a7b009eb58f8c9f256e6597711a77ded11e15d5865a3fe16901603264e1f70",
+      "size_bytes": 64953344
     },
     {
-      "current_version": "1.8.13-alpha",
-      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.13-alpha/archipelago-frontend-1.8.13-alpha.tar.gz",
-      "name": "archipelago-frontend-1.8.13-alpha.tar.gz",
-      "new_version": "1.8.13-alpha",
-      "sha256": "d0159b61f84eb30013634a97177801474376dddb1189024b94f16236ce7ff538",
-      "size_bytes": 97915796
+      "current_version": "1.8.17-alpha",
+      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.17-alpha/archipelago-frontend-1.8.17-alpha.tar.gz",
+      "name": "archipelago-frontend-1.8.17-alpha.tar.gz",
+      "new_version": "1.8.17-alpha",
+      "sha256": "faf692e9a0e16268357bcac2bf86b62950ae49663e3c95982e54a132bb761980",
+      "size_bytes": 98801608
     }
   ],
-  "release_date": "2026-09-12",
-  "signature": "45a322e793a2fc7565cc148f14efbbcf03e0c934183590375368091368ffe49ed2a7e9962692caf54b3432c71976a6c1b0f06a48cf8c4634c7b2e573e5782a0f",
+  "release_date": "2026-09-15",
+  "signature": "c8196fe278a5747b3c3ba3bf70998874f1e3e6eedbdab33b9e33c3339a3769ab4431f41d99924ec4cdd15a5ffed299a5af786c7ab5e9d084cdc11beabbee9103",
   "signed_by": "did:key:z6Mkfu5LT8d4DjETtrkATvHh9Dvcbnr7zBCUwfau8Sw7DLWT",
-  "version": "1.8.13-alpha"
+  "version": "1.8.17-alpha"
 }

releases/manifest.json

@@ -1,29 +1,30 @@
 {
   "changelog": [
-    "**GitWorkshop installs reliably on fresh nodes.** The app is classified as a user-facing app while its install placeholder is being created, so it remains visible under My Apps instead of Services.",
-    "**Fresh GitWorkshop installs build the correct image.** The production orchestrator handles its bundled build context instead of sending the local image reference through the legacy registry-pull path.",
-    "**Curated app classification is regression-tested.** Every user-facing app remains in My Apps during installation, while headless services stay in Services."
+    "Minibits claims that every mint reports as already spent leave the retry queue, clearing repeated failure notices. Network errors and mixed mint failures remain queued for another attempt.",
+    "Minibits polls its primary relay first and connects to public fallback relays only when the primary is unreachable, reducing unnecessary connections.",
+    "Large payment backlogs are fetched from newest to oldest with a saved cursor, so polling can resume after interruptions or page limits. Payments sharing the same timestamp remain reachable.",
+    "Added regression coverage for spent-claim classification, wrapped and mixed mint errors, same-second payments, and interrupted or multi-poll backlogs."
   ],
   "components": [
     {
-      "current_version": "1.8.13-alpha",
-      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.13-alpha/archipelago",
+      "current_version": "1.8.17-alpha",
+      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.17-alpha/archipelago",
       "name": "archipelago",
-      "new_version": "1.8.13-alpha",
-      "sha256": "832c7e75b395f94f919a21d1ad7d32d367e1ef84a0e0995b4e8fe87268aa21a4",
-      "size_bytes": 64585424
+      "new_version": "1.8.17-alpha",
+      "sha256": "32a7b009eb58f8c9f256e6597711a77ded11e15d5865a3fe16901603264e1f70",
+      "size_bytes": 64953344
     },
     {
-      "current_version": "1.8.13-alpha",
-      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.13-alpha/archipelago-frontend-1.8.13-alpha.tar.gz",
-      "name": "archipelago-frontend-1.8.13-alpha.tar.gz",
-      "new_version": "1.8.13-alpha",
-      "sha256": "d0159b61f84eb30013634a97177801474376dddb1189024b94f16236ce7ff538",
-      "size_bytes": 97915796
+      "current_version": "1.8.17-alpha",
+      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.17-alpha/archipelago-frontend-1.8.17-alpha.tar.gz",
+      "name": "archipelago-frontend-1.8.17-alpha.tar.gz",
+      "new_version": "1.8.17-alpha",
+      "sha256": "faf692e9a0e16268357bcac2bf86b62950ae49663e3c95982e54a132bb761980",
+      "size_bytes": 98801608
     }
   ],
-  "release_date": "2026-09-12",
-  "signature": "45a322e793a2fc7565cc148f14efbbcf03e0c934183590375368091368ffe49ed2a7e9962692caf54b3432c71976a6c1b0f06a48cf8c4634c7b2e573e5782a0f",
+  "release_date": "2026-09-15",
+  "signature": "c8196fe278a5747b3c3ba3bf70998874f1e3e6eedbdab33b9e33c3339a3769ab4431f41d99924ec4cdd15a5ffed299a5af786c7ab5e9d084cdc11beabbee9103",
   "signed_by": "did:key:z6Mkfu5LT8d4DjETtrkATvHh9Dvcbnr7zBCUwfau8Sw7DLWT",
-  "version": "1.8.13-alpha"
+  "version": "1.8.17-alpha"
 }

releases/pending/v1.8.14-alpha/manifest.json

@@ -1,30 +0,0 @@
-{
-  "changelog": [
-    "**Cuprate gains a first-party companion dashboard.** The Monero node now has a Bitcoin-style status UI, safe app grouping, a 450 GB disk-safety gate, and a restricted RPC that is never exposed as a launch page.",
-    "**Bitcoin Core Tor enrollment uses the correct protocol identity.** `bitcoin-core` is forwarded on port 8333 and resolves to its own hidden-service directory without disturbing legacy Bitcoin aliases.",
-    "**GitWorkshop opens Archipelago’s canonical ngit repository by default.** The launcher and registry promotion use the full maintainer/relay/`archy` coordinate, with regression coverage for Companion and browser-tab launches.",
-    "**Release validation is stricter.** The registry gate now checks the complete canonical source deep link, and the merged candidate passed the full frontend and focused backend test suites."
-  ],
-  "components": [
-    {
-      "current_version": "1.8.14-alpha",
-      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.14-alpha/archipelago",
-      "name": "archipelago",
-      "new_version": "1.8.14-alpha",
-      "sha256": "3d8e5e7c79a261649e89c4f5ba8d90db9057ebbb002919a812ca82f664b315bf",
-      "size_bytes": 64568360
-    },
-    {
-      "current_version": "1.8.14-alpha",
-      "download_url": "https://source.archipelago-foundation.org/lfg2025/archy/releases/download/v1.8.14-alpha/archipelago-frontend-1.8.14-alpha.tar.gz",
-      "name": "archipelago-frontend-1.8.14-alpha.tar.gz",
-      "new_version": "1.8.14-alpha",
-      "sha256": "e1c490e52571bf9238435e5986792c6bd602fd7e398783546f7592aa921d3386",
-      "size_bytes": 98792963
-    }
-  ],
-  "release_date": "2026-09-13",
-  "signature": "dacfdd2707e415af8a42306a63658a7d41cdfeba29d43802d465d18f00a747ecbfa54ea4ee8ed1eedf94d4f30371453fd9c9d475af9d6a62eac4e2c8e783b405",
-  "signed_by": "did:key:z6Mkfu5LT8d4DjETtrkATvHh9Dvcbnr7zBCUwfau8Sw7DLWT",
-  "version": "1.8.14-alpha"
-}

scripts/image-versions.sh

@@ -33,7 +33,8 @@ ELECTRUMX_IMAGE="$ARCHY_REGISTRY/electrumx:v1.18.0"
 
 # Mempool stack
 MEMPOOL_BACKEND_IMAGE="$ARCHY_REGISTRY/mempool-backend:v3.3.1"
-MEMPOOL_WEB_IMAGE="$ARCHY_REGISTRY/mempool-frontend:v3.3.1"
+# The patched frontend is published by chaum on the same trusted registry.
+MEMPOOL_WEB_IMAGE="source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1"
 MARIADB_IMAGE="$ARCHY_REGISTRY/mariadb:11.4.10"
 
 # BTCPay

scripts/publish-release-assets.sh

@@ -53,24 +53,33 @@ if [ -x "$PROJECT_ROOT/core/target/release/archipelago" ]; then
 fi
 
 remote_url=$(git -C "$PROJECT_ROOT" remote get-url "$REMOTE")
-# https is accepted as well as http. Requiring http:// meant the only remote
-# whose credential actually works for git push (the https one) was rejected,
-# while the http remote it forced you to use had a dead token — so publishing
-# failed on auth after the manifest had already passed every check
-# (v1.7.121-alpha, 2026-08-04). The scheme is carried through to the API URL
-# rather than assumed.
+# Remote URLs are public metadata: ngit can include them in repository
+# announcements. Keep credentials in Git's credential helper, never in URLs.
 case "$remote_url" in
-    http://*@*|https://*@*) ;;
-    *) fail "$REMOTE must be an authenticated http(s):// Gitea remote URL for API uploads" ;;
+    http://*@*|https://*@*) fail "$REMOTE embeds credentials; move them to a Git credential helper and remove them from the remote URL" ;;
+    http://*|https://*) ;;
+    *) fail "$REMOTE must be an http(s):// Gitea remote URL for API uploads" ;;
 esac
 
 scheme=${remote_url%%://*}
-rest=${remote_url#*://}
-auth=${rest%%@*}
-host_path=${rest#*@}
+host_path=${remote_url#*://}
 host=${host_path%%/*}
 repo_path=${host_path#*/}
 repo_path=${repo_path%.git}
+credential=$(printf 'url=%s\n\n' "$remote_url" | GIT_TERMINAL_PROMPT=0 git -C "$PROJECT_ROOT" credential fill) \
+    || fail "no Git credential available for $REMOTE; configure a credential helper first"
+auth_user=""
+auth_password=""
+while IFS= read -r field; do
+    case "$field" in
+        username=*) auth_user=${field#username=} ;;
+        password=*) auth_password=${field#password=} ;;
+    esac
+done <<< "$credential"
+[ -n "$auth_user" ] && [ -n "$auth_password" ] \
+    || fail "Git credential helper did not provide a username and password for $REMOTE"
+auth="$auth_user:$auth_password"
+unset credential auth_user auth_password
 api="$scheme://$host/api/v1/repos/$repo_path"
 release_url="$api/releases/tags/v${VERSION}"
 

scripts/test-mempool-dns-recovery.py

@@ -0,0 +1,136 @@
+#!/usr/bin/env python3
+"""Exercise the built frontend against a backend that disappears and changes IP.
+
+Uses an isolated Podman network and disposable containers, never the node stack.
+Usage: python3 scripts/test-mempool-dns-recovery.py [frontend-image]
+"""
+import ipaddress
+import json
+import socket
+import subprocess
+import sys
+import time
+import urllib.error
+import urllib.request
+import uuid
+
+IMAGE = sys.argv[1] if len(sys.argv) > 1 else (
+    "source.archipelago-foundation.org/chaum/mempool-frontend:v3.3.1-archy1"
+)
+BACKEND = "source.archipelago-foundation.org/lfg2025/mempool-backend:v3.3.1"
+prefix = "mempool-dns-test-" + uuid.uuid4().hex[:8]
+network, frontend, backend = prefix, prefix + "-web", prefix + "-api"
+
+
+def podman(*args, check=True):
+    return subprocess.run(["podman", *args], capture_output=True, text=True,
+                          check=check, timeout=60).stdout.strip()
+
+
+def eventually(check, timeout=25):
+    deadline = time.monotonic() + timeout
+    while True:
+        try:
+            return check()
+        except (AssertionError, OSError, urllib.error.URLError):
+            if time.monotonic() >= deadline:
+                raise
+            time.sleep(1)
+
+
+server = r"""
+const http = require('http'), crypto = require('crypto');
+const server = http.createServer((req, res) => {
+  res.setHeader('Content-Type', 'application/json');
+  res.end(JSON.stringify({url: req.url, instance: process.env.INSTANCE}));
+});
+server.on('upgrade', (req, socket) => {
+  const key = crypto.createHash('sha1')
+    .update(req.headers['sec-websocket-key'] + '258EAFA5-E914-47DA-95CA-C5AB0DC85B11')
+    .digest('base64');
+  socket.end('HTTP/1.1 101 Switching Protocols\r\nUpgrade: websocket\r\n' +
+    'Connection: Upgrade\r\nSec-WebSocket-Accept: ' + key + '\r\n' +
+    'X-Upstream-Url: ' + req.url + '\r\nX-Instance: ' + process.env.INSTANCE + '\r\n\r\n');
+});
+server.listen(8999, '0.0.0.0');
+"""
+
+try:
+    podman("network", "create", network)
+    subnet = ipaddress.ip_network(json.loads(podman("network", "inspect", network))[0]["subnets"][0]["subnet"])
+    podman("run", "-d", "--name", frontend, "--network", network,
+           "-p", "127.0.0.1::8080", "-e", "BACKEND_MAINNET_HTTP_HOST=mempool-api",
+           "-e", "FRONTEND_HTTP_PORT=8080", IMAGE)
+    port = int(podman("port", frontend, "8080/tcp").rsplit(":", 1)[1])
+    url = f"http://127.0.0.1:{port}"
+
+    def static_ready():
+        assert urllib.request.urlopen(url, timeout=4).status == 200
+
+    eventually(static_ready)
+    started = podman("inspect", frontend, "--format", "{{.State.StartedAt}}")
+    try:
+        urllib.request.urlopen(url + "/api/v1/backend-info", timeout=6)
+        raise AssertionError("An absent backend must not appear healthy")
+    except urllib.error.HTTPError as error:
+        assert error.code == 502
+    print("PASS: frontend starts while backend DNS is absent", flush=True)
+
+    for instance, offset in [("first", 10), ("replacement", 11)]:
+        if instance == "replacement":
+            podman("rm", "-f", backend)
+            # Ensure the cached address has expired while the backend is absent.
+            time.sleep(6)
+        podman("run", "-d", "--name", backend, "--network", network,
+               "--network-alias", "mempool-api", "--ip", str(subnet[offset]),
+               "-e", "INSTANCE=" + instance, "--entrypoint", "node", BACKEND,
+               "-e", server)
+
+        for path, expected in [
+            ("/api/blocks/tip/height?probe=one", "/api/v1/blocks/tip/height?probe=one"),
+            ("/api/v1/fees/recommended?probe=two", "/api/v1/fees/recommended?probe=two"),
+        ]:
+            def check_http():
+                with urllib.request.urlopen(url + path, timeout=4) as response:
+                    result = json.load(response)
+                assert result == {"url": expected, "instance": instance}, result
+            eventually(check_http)
+
+        for path in ["/api/v1/ws?probe=ws", "/ws?probe=ws"]:
+            def check_ws():
+                with socket.create_connection(("127.0.0.1", port), timeout=4) as sock:
+                    sock.sendall((f"GET {path} HTTP/1.1\r\nHost: localhost\r\n"
+                                  "Upgrade: websocket\r\nConnection: Upgrade\r\n"
+                                  "Sec-WebSocket-Version: 13\r\n"
+                                  "Sec-WebSocket-Key: dGhlIHNhbXBsZSBub25jZQ==\r\n\r\n").encode())
+                    response = b""
+                    while b"\r\n\r\n" not in response:
+                        part = sock.recv(4096)
+                        assert part, response
+                        response += part
+                assert b"101 Switching Protocols" in response, response
+                assert b"X-Upstream-Url: /?probe=ws" in response, response
+                assert ("X-Instance: " + instance).encode() in response, response
+            eventually(check_ws)
+        assert podman("inspect", frontend, "--format", "{{.State.StartedAt}}") == started
+        print(f"PASS: {instance} backend at {subnet[offset]}: HTTP paths, query strings, both WebSocket routes; frontend never restarted", flush=True)
+
+    before = podman("exec", frontend, "cat", "/etc/nginx/conf.d/nginx-mempool.conf")
+    podman("exec", frontend, "/patch/repair-nginx.sh")
+    assert podman("exec", frontend, "cat", "/etc/nginx/conf.d/nginx-mempool.conf") == before
+    podman("exec", frontend, "nginx", "-t")
+    print("PASS: repeated repair is idempotent and nginx configuration is valid", flush=True)
+    podman("restart", frontend)
+    eventually(static_ready)
+
+    def after_restart():
+        with urllib.request.urlopen(url + "/api/blocks/tip/height?restart=1", timeout=4) as response:
+            assert json.load(response) == {
+                "url": "/api/v1/blocks/tip/height?restart=1", "instance": "replacement"
+            }
+    eventually(after_restart)
+    eventually(check_ws)
+    print("PASS: frontend restart preserves DNS recovery and HTTP/WebSocket routing", flush=True)
+finally:
+    podman("rm", "-f", frontend, backend, check=False)
+    podman("network", "rm", network, check=False)

tests/release/run.sh

@@ -168,7 +168,7 @@ stage "cargo-check"      timeout 580 cargo check --manifest-path core/Cargo.toml
 # 3600s leaves headroom; a warm target/ finishes in a fraction of it.
 stage "cargo-test-weekly" timeout 3600 env CARGO_INCREMENTAL=0 \
   cargo test --manifest-path core/Cargo.toml -p archipelago -- \
-  update:: lnd container::image_versions scanner drift missing_secret collision
+  update:: lnd container::image_versions upgrade_preserves_container scanner drift missing_secret collision
 
 # ── Stage 4: live node smoke ─────────────────────────────────────────
 if [[ $LIVE -eq 1 ]]; then