Sizes bitcoind -dbcache to host RAM (~1/16, floor 300MB, cap 4096) instead of a
fixed 2048/4096. A multi-GB UTXO cache on an 8GB node running the full app stack
pushed memory past physical RAM and triggered system-wide swap thrash: the disk
saturated, bitcoind could not answer its own RPC, and the dashboard backend's
sqlite reads stalled — surfacing as fleet-wide /rpc/v1 502s and a blank Bitcoin
UI. Applied in scripts/container-specs.sh (reconciler path) and the config.rs
bitcoin-core path.
Bitcoin status cache now polls every 5s (was 10/15) with an 8s timeout (was 20s)
and fetches the four RPCs concurrently, so the cached snapshot tracks bitcoind's
responsive windows during IBD and the UI stops dwelling on "reconnecting...".
Unifies the divergent discover AppGrid/FeaturedApps image-error handlers onto the
canonical placeholder fallback so missing app icons render the placeholder.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- DMs now use native meshcore unicast (CMD_SEND_TXT_MSG) instead of @DM2 channel
broadcasts: private (E2E-encrypted to the recipient pubkey by firmware), off the
public channel, and decodable by stock clients. Plain text (split, not MC-chunked)
to non-archipelago contacts; typed envelopes to archy peers.
- !ai replies now DM the asker privately (RadioDm) instead of broadcasting on ch0.
- Auto contact-import: a heard advert (PUSH_CONTACT_ADVERT/0x80, 32-byte pubkey) is
added via CMD_ADD_UPDATE_CONTACT (0x09) so contacts appear without a flood advert.
- clear-all now DELETES firmware contacts via CMD_REMOVE_CONTACT (0x0F) instead of
blocklisting; blocking filter removed entirely. Wiped contacts return when reachable.
- Contact reachability: MeshPeer carries last_advert + reachable (path-based); UI shows
a reachability dot.
- Peers list: contact search box (filter by name/DID/npub/pubkey) with a clear button.
- send_message routes stock contacts as plain native text (fixes garbled envelopes).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- mesh: stop broadcasting ARCHY:2 identity on the public channel (startup + every advert tick); receive path still parses inbound. No more public-channel spam.
- mesh assistant: trigger on !ai/!ask typed in 1:1 chat (was only the dead AssistQuery path + bare channel text); route the reply transport-aware via MeshService::send_message (Tor for federation peers, LoRa for radio) through a new AssistChatReply event consumed at the server layer — fixes replies never reaching federation askers.
- mesh assistant: per-contact !ai allowlist (allowed_contacts) bypassing trusted_only; config + RPC + is_sender_allowed.
- fedimint-clientd manifest: network_policy open -> bridge (invalid value made the loader skip the whole manifest, so fmcd never ran and federations never joined/listed).
- ui: AI panel — Claude model dropdown (Haiku/Sonnet/Opus presets) + allowlist contact picker.
- ui: Settings — App Updates + App Registry moved under Account.
- ui: mesh chat — overscroll-behavior: contain so chat scroll no longer bleeds to the contacts panel.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Component tests mounted without main.ts's bootstrap, so the $ver global
template helper (app.config.globalProperties.$ver = displayVersion) was
undefined — AppSidebar/AppHeroSection/MarketplaceAppCard tests failed with
"_ctx.$ver is not a function", blocking the release gate's ui-unit-tests
stage. Add a vitest setup file that mirrors main.ts via config.global.mocks
and wire it into vitest.config.ts.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds the assistant scheduler, MeshAssistantPanel UI, and the remaining
config-RPC / live-toggle / Ollama-detect wiring on top of Phase 1.x.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Phase 2 backend. AssistantConfig is now live-updatable (RwLock) so the UI
toggle applies without a listener restart. New RPCs:
- mesh.assistant-status -> {enabled, model, trusted_only, default_model,
ollama_detected, models[]} (probes local Ollama :11434/api/tags)
- mesh.assistant-configure -> set enabled/model/trusted_only live + persist
MeshService::assistant_config / configure_assistant. Compiles clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The ARCHY:2 identity broadcast (DID + ed25519 + x25519) was unwired dead
code on both send and receive. Wiring it lets a radio peer prove its
archipelago identity, so the assistant's trusted-only gate (and encrypted
DMs) work over meshcore AND Meshtastic — the latter otherwise only exposes
synthetic node keys.
- session.rs: broadcast ARCHY:2 as channel text at startup + each advert tick
- frames.rs: parse inbound ARCHY:2 on the channel path, dedupe-keyed by
archipelago pubkey (federation_peer_contact_id) so it MERGES with the
federation-seeded peer instead of duplicating; self-echo guarded
- threads our_x25519_secret into handle_channel_payload (was reserved)
Reuses the existing handle_identity_received verifier (ed/x25519 consistency
check + shared-secret derivation). Compiles clean. Needs a live 2-radio test
before trusting trusted-only over radio.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A plain '!ai <q>' / '!ask <q>' on the channel is now answered by the node's
local model and broadcast back as plain text, so ANY client (bare meshcore
or Meshtastic) can ask. Generalised run_assist with an AssistReply target:
Typed chunks to a peer (archipelago UI path) vs plain channel-text (bare
clients). Trust/rate gate unchanged; asker identity is separate from reply
mode. Works over both radios.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Headless containers (databases, APIs, backends without a UI) belong in a
tab labelled 'Services', not 'Websites'. The categorisation logic already
routes UI-less packages there (built under #45); this finishes the rename
of the user-facing label across Apps, Marketplace, Discover and the mobile
nav, and makes 'services' the canonical tab state/query param. Old
?tab=websites bookmarks still resolve (back-compat acceptor kept).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Destructure the first 4 pubkey bytes into typed locals so vue-tsc's
noUncheckedIndexedAccess doesn't fail the build (the bytes.length<4 guard
doesn't narrow per-element access). No behaviour change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The IndeeHub API needs MinIO (object storage) up to serve, but the
health monitor's dependency map listed only postgres + redis, so it
would restart the API while MinIO was still starting — the "recovers
only after 1-2 container restarts" symptom. Add indeedhub-minio to the
API's deps; MinIO has no deps of its own so the monitor restarts it
first, no deadlock. (First-start ordering in the stack definition is a
deeper, separate follow-up.)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
federation.remove-node only edited nodes.json, so a removed/renamed node
(e.g. a stale "Arch HP") lingered in the mesh chat list with its old
thread. Capture the node's pubkey before removal, then purge its
synthetic mesh peer, shared secret, messages, presence, and persisted
contact entry via the new mesh::purge_federation_peer. Combined with the
#42 name refresh, stale federation contacts can now be fully cleaned from
a node.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Per the rule that only front-end apps with a UI belong in "My Apps"
(databases/backends/headless go to Websites), make the manifest's
interfaces.main.ui the deciding signal. isWebsitePackage now treats any
package that declares a UI as an app even when it isn't in the curated
APP_CATEGORY_MAP, and falls through headless LAN-reachable packages to
Websites. Additive — service-by-name infra and curated known apps are
unchanged, so no currently-correct app moves.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Settings shows the node-level Nostr key (HKDF derive_node_nostr_key,
read via node.nostr-pubkey) while Web5 > Identities showed the identity
record's own key — the mirrored "Node" identity stores nostr=None and
seed identities use a different BIP-32 NIP-06 key, so the two surfaces
disagreed.
Resolve the node-level Nostr key once in identity.list and override it
onto whichever identity record is the node's own (ed25519 == server_info
.pubkey). Display-only — no stored key is rewritten, so it self-applies
to existing nodes with no migration and the discovery identity is
unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A peer accepted via invite is seeded into the mesh peer table with
name=None, so it shows as "Archipelago <pubkey8>" in chat. Federation
sync later learns the real name (update_node_state writes it to
nodes.json) and discovers transitive peers (merge_transitive_peers),
but nothing pushed those into the live mesh peer table — the chat list
stayed stale until the next mesh restart, and transitive peers never
appeared as contacts at all.
Add RpcHandler::refresh_federation_mesh_peers() (re-runs the idempotent,
onion-deduped seed_federation_peers_into_mesh) and call it after every
periodic sync cycle (server.rs) and after the manual federation.sync-all
RPC. Names now correct themselves and the full roster meshes within a
sync cycle, no restart needed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
phase4-streaming-ecash-plan.md: design for ecash-paid swarm transport, paying
across different mints (§2a, Lightning-bridged swaps), networking-through-nodes
relay, and an IndeeHub "Archipelago" content source. Records the resolved
iroh-blobs paid-serving spike. dht-RESUME.md: task #12 + step F marked done.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds a Settings control to the Networking Profits card that opens a new page
where the operator controls what their node charges sats for and how much.
Drives the existing streaming.list-services / streaming.configure-service RPCs;
"free everything" is the default (all priced services ship disabled, surfaced
with a reassurance banner). New route web5/networking-profits + common.settings
i18n (en/es).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Phase 3 wiring (task #12):
- NostrSeedDiscovery: async ProviderDiscovery that queries relays for signed
seed adverts and parses endpoint ids (swarm/iroh_provider.rs, seed_advert.rs).
- seed_and_advertise publish path; dep-free fetch/publish helpers reuse the
node's Nostr identity (build_nostr_client/load_or_create_nostr_keys made
pub(crate)).
- swarm::init builds the IrohProvider once into a OnceLock runtime; providers()
returns it; announce_held_blob() is called from update.rs after a release
component passes both hash gates.
- config swarm_enabled (ARCHIPELAGO_SWARM_ENABLED, default off); server.rs init.
Paid swarm serving (Phase 4 step F):
- swarm/paid.rs gates the iroh-blobs provider through streaming::gate,
intercepting connect + GET (peer push hard-disabled). Free by default
(content-download service disabled); denies unpaid peers when enabled;
fails open on internal error so a payment fault never blocks distribution.
Wired into IrohProvider::new.
All iroh code behind the iroh-swarm feature; the default build is inert.
Default build clean; --features iroh-swarm: 11/11 swarm tests pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
bitcoin-core was missing from APP_CATEGORY_MAP, so isKnownApp() was false and
isWebsitePackage() fell through to 'has a runtime LAN address'. Once the running
container's LAN address (the bitcoind RPC port :8332) showed up ~a minute after
launch, Bitcoin Core was reclassified as a website: it dropped out of the Apps
tab and search, moved under Websites, and launching it opened :8332 (raw RPC)
instead of the :8334 custom UI that Knots opens.
Add 'bitcoin-core': 'money' alongside bitcoin-knots/bitcoin-ui so isKnownApp is
true, isWebsitePackage is false, and launchAppNow routes through openSession ->
resolveAppUrl (:8334 custom UI). Fixes search, category, and the launch URL.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Messaging a federation-only peer (e.g. 'Arch Dev') failed with 'Missing
contact_id'. The UI gave federation-only rows a *negative* placeholder
contact_id derived from a DID hash, but the backend parses contact_id as u64,
so a negative value deserialized to None. The negative id also never matched
the positive federation-synthetic id that federation-routed messages are stored
under, so those threads looked empty.
- Frontend: derive the SAME positive federation-synthetic id the backend uses
(federationContactId mirrors federation_peer_contact_id) so mesh.send accepts
it and messages thread correctly.
- Backend: send_typed_wire now resolves a federation-synthetic contact_id from
nodes.json when it isn't in the live mesh peer table (radio-less node),
instead of bailing 'Unknown federation peer'.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Large peer downloads (~178MB) failed with a generic 'Operation failed', and
the download path had three stacked problems:
- The FIPS reqwest client used a hard-coded 20s total timeout regardless of the
caller's .timeout(), so a big transfer over the mesh aborted at 20s before
the Tor fallback could help. Honor the per-request timeout (client_with_timeout).
- The peer-content proxy buffered the whole file into node memory via
resp.bytes() before sending a byte, and capped the transfer at 60s. Stream
the body through with hyper::Body::wrap_stream (constant memory) and raise the
timeout to 900s; bump the nginx peer-content read timeout to match.
- Free downloads pulled the file as base64 over RPC, doubling it in node memory
and the browser — fatal for large files. Download free files by streaming
from /api/peer-content straight to disk, after a 1-byte Range probe that
surfaces the real reason (peer offline on mesh and Tor) instead of a generic
failure. Paid downloads now return the real error through the {error} channel
the UI already displays.
Adds the reqwest 'stream' feature for bytes_stream().
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The prod orchestrator only checked whether a build-image tag was *present*
before deciding to skip the build. The local UI images (bitcoin-ui, lnd-ui,
electrs-ui) COPY a built neode-ui dist, so a UI update changed the source but
left the old tag in place and the new UI never shipped.
Gate the build on a content fingerprint of the build context (sorted relative
path + length + mtime, SHA-256) recorded in a per-tag stamp under data_dir.
Rebuild whenever the fingerprint differs from the one that produced the
existing image; podman's own COPY-layer cache keeps a no-op rebuild cheap.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Single source of truth for picking the DHT work back up after a restart:
worktree/branch rules, all phase commits, the exact next task (#12 Phase 3
glue), build-time facts, and the Phase 0 go-live ceremony.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Apps could fail install when a stack member exited on its first start
because a dependency (db/redis/the bitcoin node) was not ready yet — a
transient crash, not a broken install. wait_for_stack_containers now
restarts each exited/dead container up to 3 times before declaring the
install failed; the runtime supervisor keeps it alive afterwards.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The discovery wire format that feeds the swarm's ProviderDiscovery seam: a
node announces 'I seed blake3 H from iroh endpoint E' as a signed NIP-33
addressable Nostr event. Scope is releases/catalog content ONLY (decided
2026-06-16) — never private user blobs.
- swarm/seed_advert.rs: kind 30081, d-tag = blake3 hex (one current advert
per author+hash, latest-replaces), content {"v":1,"endpoint_id":...}.
advertisement_builder / advertisement_filter / parse_endpoint_id /
endpoint_ids_from_events (dedup). Endpoint ids stay opaque strings so the
protocol is dep-light + unit-testable on the default build.
4/4 tests pass (sign->parse roundtrip, filter targeting, reject wrong-kind/
empty, dedup across nodes).
Next (task #12): gated NostrSeedDiscovery glue (query relays, parse ids ->
iroh::EndpointId), publish path, wire swarm::providers().
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Pulls iroh 1.0 + iroh-blobs 0.103 as OPTIONAL deps under the iroh-swarm
feature and implements a real BlobProvider over them. Verified: the full
iroh QUIC dep tree (260 pkgs) resolves and compiles against the pinned
bitcoin/nostr-sdk/reqwest-rustls stack; the provider compiles against the
0.103/1.0 API.
- swarm/iroh_provider.rs: IrohProvider::new binds a QUIC Endpoint, opens a
persistent FsStore (data_dir/iroh-blobs), and serves blobs via the
iroh-blobs protocol/Router — a node that fetches also SEEDS. try_fetch
maps ContentDigest -> iroh Hash, asks discovery for seed EndpointIds, then
downloader.download(hash, providers) (range-verified) + export to staging.
- ProviderDiscovery trait: the seam Phase 3 (signed Nostr advertisement
events) fills. discovery=None -> no seeds -> origin-only, so enabling the
feature is never worse than today.
- Default build untouched: iroh is optional, the module is cfg-gated, and
providers() stays empty until Phase 3 wires discovery in.
Build: cargo build --features iroh-swarm succeeds (dev). Default build +
44 swarm/update/content_hash/blobs tests unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The backend already sends did in federation peer lists, but the Peer
type omitted it and federationNodeToPeer() dropped it when mapping. Add
did?: string to Peer and pass node.did through, so trusted/observer
node rows route to Federation/Mesh by their real DID (falling back to
pubkey/onion) instead of failing the build on a missing property.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Lands the transport/swarm orchestration layer (the iroh engine attaches
later, behind a flag). The seam is fully exercised today with the origin
HTTP path; with no swarm providers registered the behaviour is byte-for-byte
identical to before.
- swarm/mod.rs: BlobProvider trait + fetch_content_addressed() — tries each
provider in order, VERIFIES peer-sourced bytes against the content digest
before accepting (untrusted seeds can't inject tampered bytes), falls back
to the origin closure if none serve. Returns Swarm|Origin.
- Cargo: iroh-swarm feature (off by default; heavy QUIC dep tree attaches
here). providers() is empty until enabled → every fetch hits origin.
- update.rs: components with a BLAKE3 digest route through the seam, using
the existing resumable HTTP downloader as the origin fallback; a swarm hit
is re-checked against the mandatory SHA-256 manifest gate (re-fetch from
origin on any disagreement). Components without blake3 take the original
path untouched.
44/44 swarm/update/content_hash/blobs tests pass (incl. swarm hit/miss,
tampered-bytes-rejected→origin, fall-through ordering).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds the iroh-native, range-verifiable hash next to the incumbent SHA-256
so the swarm can later fetch/verify by BLAKE3 with the registry/origin as
fallback. Non-breaking: SHA-256 stays the mandatory gate; BLAKE3 is verified
only when present.
- content_hash.rs: HashAlg + ContentDigest (parse/verify '<alg>:<hex>'
multihash strings), blake3_hex/sha256_hex; BLAKE3 known-answer test
- update.rs: ComponentUpdate.blake3 (serde-default); verified ALONGSIDE
SHA-256 in the resumable download loop, re-download on mismatch
- blobs.rs: BlobMeta.blake3 computed on put (on-disk path stays
SHA-256-keyed for back-compat; advertises the future swarm address)
Drive-by: fix a pre-existing stale test (test_save_and_load_state_roundtrip)
that never wrote the .download-complete marker #26 requires, so load_state's
self-heal cleared update_in_progress. Unrelated to BLAKE3 — surfaced by
running the full update:: suite.
40/40 content_hash/update/blobs tests pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The #26 fix makes has_staged_update require the .download-complete
marker, so the state self-heal treats a marker-less staging dir as a
partial download and clears update_in_progress. The roundtrip test
staged a binary file but not the marker, so it began failing. Write
the marker to simulate a *complete* staged update.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Completes the parked trust module and wires it into the live build:
- main.rs: register `mod trust`
- app_catalog::fetch_one: verify the release-root detached signature when
present (verify against raw JSON so forward-compat fields stay in the
signed preimage); accept unsigned during the migration window, hard-reject
a present-but-bad signature so a tampering mirror can't pass altered bytes
- seed: pin release-root Ed25519 known-answer test (priv+pub) for the
signing ceremony / pinned-anchor / external-verifier cross-check
- signed_doc: drop unused import
20/20 Phase 0 unit tests pass (trust::canonical/did/signed_doc/anchor,
seed release-root, app_catalog). Crate compiles clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
