Commit Graph
38 Commits
Author SHA1 Message Date
Kgothatso NgakoandClaude Opus 5.5 3393cd2b3f docs: plan a group Lightning wallet held with an Iceberg key
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 16:10:08 +02:00
Kgothatso Ngako 409be6db28 Merge remote-tracking branch 'origin/bennu' into claude/satoshi-amount-group-wallet-a12572
# Conflicts:
#	composeApp/src/commonMain/kotlin/com/machankura/bennu/compose/wallet/GroupOnchainWallet.kt
#	composeApp/src/jvmTest/kotlin/com/machankura/bennu/compose/wallet/GroupOnchainWalletJvmTest.kt
#	docs/README.md
2026-10-23 12:51:52 +02:00
Kgothatso NgakoandClaude Opus 5.5 d56c41ffa4 docs(wallet): send-an-amount checked, and what is left to do by hand (Phase 6)
The audit and the JVM suite ran after every phase. The two-device regtest
walkthrough has not been done; the note lists each of its steps beside
the automated test that comes closest and what that test cannot show.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 12:34:07 +02:00
Kgothatso NgakoandClaude Opus 5.5 f9e4e049f8 docs: the group wallet's address list is built (Phase 5)
Marks the plan built, with a table of the nine places the build chose
differently, and points the group-chain-wallet note's gap-limit
paragraph and the docs index at it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 12:26:59 +02:00
Kgothatso NgakoandClaude Opus 5.5 bc348419bc feat(wallet): read an amount as whole sats, or say why not (Phase 1)
The gated amount field read its text by dropping every non-digit, so
21000.5 became 210,005 sats and 0.0005 became 5. parseSats accepts
thousands grouping and refuses decimals, grouping that groups nothing,
non-ASCII digits and more sats than exist; the preview reports which.
The field shows plain digits grouped and keeps what was typed.

Also plans the rest in docs/group-send-an-amount.md: send-an-amount
opens on every network at once, held on mainnet by the existing cap.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 12:15:17 +02:00
Kgothatso NgakoandClaude Opus 5.5 6bf90d9d70 docs: plan a list of the group wallet's addresses
Modelled on Phoenix's swap-in addresses screen: every paid receive
address, the current one, and the two after it, read off the receive
mini-wallet's state with nothing stored and nothing new subscribed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 12:13:37 +02:00
Kgothatso NgakoandClaude Opus 5.5 f83fe35a45 feat(wallet): ship behind the rollout gates (Phase 10)
GroupWalletRollout holds the three gates, each with what lifts it. The
flag is on, its condition met. Send-an-amount is off, so the form sends
everything confirmed to one address and says so. A spend proposed from
this device on mainnet may send at most 100,000 sats; test networks
have no cap. The gates are on the composer and nothing else: a device
behind them still accepts, shows and signs what another member
proposes, through every check GroupSpend makes.

The plan is marked built, with its table of where the build chose
differently. Full JVM suite: 1474 tests, none failing; the Android
target compiles; every M3 budget is met.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 08:52:16 +02:00
Kgothatso NgakoandClaude Opus 5.5 dbbc6c6474 docs: plan a group chain wallet on the group's threshold key
Modelled on Phoenix's final wallet, phased. The path is BIP86's
unhardened tail under a public account step over the ChillDKG key, so
the group can export a standard tr() descriptor; each input is a batch
item with its own tweak cache, and a member checks seven things before
signing a spend. Shares are used as they are at rest, for now.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 07:47:25 +02:00
Kgothatso NgakoandClaude Opus 5 2edd05f3ca fix(frost): a broadcast that queued nothing records nothing
`broadcast` now says whether it queued anything, and `publishOwn` records the
local copy only when it did. The copy is what stops `advance` from publishing
that message again, so writing one for a message that never went out is a
message the group waits on forever -- the exact opposite of the trade the
queue-then-record order exists to make. The comment on `publishOwn` had that
trade written down in one direction only.

Reachable only through the id-collision guard added with the resend, which
refuses after four tries rather than overwriting a sent row. Four consecutive
seconds of the same message is not something a person with a button can do, but
"unreachable" and "safe when reached" are different claims.

Also brings the docs up to what was built: the resend note is in the index, the
skipped-keys note points at it as the recovery for the failure it describes --
a new message is encrypted at a generation the receiver has not passed, which is
the whole mechanism -- and the plan records the two things landing it turned up,
plus `ask_again` being reused from the catalogue rather than duplicated.

1279 tests, no failures. m3Audit: all budgets met.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 15:09:57 -05:00
Kgothatso NgakoandClaude Opus 5 de5b2bd4a6 docs(marmot): the v2 plan is built -- where the build chose differently, and where the rollout stands (v2 Phase 9)
docs/marmot-v2.md, Phase 9. The phase has no code in it; it is the rollout,
and the rollout is a flip and a retirement that wait on the people this
app's users invite. What the phase owes the document is the repo's
convention for a plan that has been built: the status paragraph, the table
of where the implementation chose differently, the thing found on the way
that belongs to nobody's phase, and the commit the account was written
against. All four are here now, and the phases themselves are kept as
written, because they are the reasoning and the code reads better against
the argument it came from than against a summary of itself.

The table has twenty-two rows, and the first is the one to read before any
of the code: the plan said the library's engine replaces the app's pipeline,
and the build made a room's profile pick its pipeline instead -- legacy
rooms keep the app's path, comparator and cache included, and retire with
the profile; current-profile rooms go through `MarmotEngine`. The library's
processor has no carve-out for the bare gift wrap a legacy direct message is
carried in, so under it every one of those would have stopped decrypting,
and the legacy pipeline is frozen anyway. Most of the other rows follow from
that one: one store rather than four, no message-id column, the rotation
manager constructed and unused, the engine hanging off the database rather
than an identity's scope. The rest record the smaller turns -- the envelope's
kind, the chronicle cap measured escaped, the KeyPackage not last-resort, the
admin refused before a leave is attempted, captions from the participants
diff rather than a snapshot column, no rename, the two non-chat kinds only,
the four Phase 8 scenarios spread over four tests -- and the three
follow-ups that closed lines the phases' own commits had missed: the
outbound hold while a pass is open, the rejected commit discarded, and the
1210/1009 bubbles.

The thing found on the way is quartz 1.15.2's: its convergence rebuilds a
candidate branch by replaying its commits against the retained parent, and
a member cannot replay a commit it authored, so a fork this device is half
of is scored with its own branch stale and two committers swap branches
rather than agree. Every device that only observed still agrees, which is
the case the app depends on and the first fork test holds; the second pins
the swap and says what to do the day it fails. The paragraph says the
practical exposure is small and why, and that the Phase 5 follow-up on
withdrawing a losing branch's rows waits on the same fix.

Phase 9's own section gains "Where it stands": everything ships built and
unflipped -- `DEFAULT_FOR_NEW_ROOMS` is legacy, both slots are published, no
room is on the current profile unless a caller asked by name -- so the first
of the three one-way doors is behind us and the other two wait on the flip.
The README's row and reading-order paragraph now say the plan has been
built and point at the table's first and last rows.

One code change, a comment: the plan's Phase 5 asked that the two library
gaps the rewrite does not close -- the retained-epoch decrypt without sender
verification, and the skipped keys that do not survive `saveState` -- be
said where the new code is, rather than let the rewrite look like it closed
them. `MarmotEngine`'s class doc now says both.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 00:16:10 +02:00
Kgothatso NgakoandClaude Opus 5 f304e7d642 docs: the Marmot v2 plan -- bringing the group layer onto the current profile
quartz 1.15.2 implements the restructured Marmot spec, and bumping the version
gives this app almost none of it: the pipeline that reads, forks, publishes and
persists group state is the app's own, driving MlsGroup directly with Room as
the store, while everything v2 lives in the library's managers behind four store
interfaces the app has never implemented. This note is what would have to change,
in the order it can safely change.

Nine rules reach the app, tabled against what it does today: unsigned six-member
payloads, component-based group state with per-leaf identity proofs, a random
routing id, dedup by the Marmot message id, convergence in place of the
created_at-then-event-id rule, publish-before-apply, SelfRemove and derived
captions, NIP-65 KeyPackage relays, no encoding tag. Two constraints shape the
phases: adopt the library's engine rather than re-implement it (convergence is
two thousand lines whose only value is agreeing with every other implementation),
and split the room's derived id -- the group's FROST key, per the derivation
note -- from the routing address v2 says must never be derived.

Nine phases, ordered so that the ones that reach other clients come before the
ones that make the app's own forks settle the way theirs do. Two findings from
the research shaped them: the build already in the field reads a canonical
payload because Event.fromJsonOrNull defaults a missing sig, so the payload
switch needs no gating; and the app picks a peer's KeyPackage by newest
created_at, so the two-slot transition has to stamp the legacy package later or
old builds break on a framed one.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 22:04:41 +02:00
Kgothatso NgakoandClaude Opus 5 b3bcc06047 docs: record what the profile preview plan built, and the places it chose differently
Two strings that already existed, a room rule that follows the old code's
intent rather than its branch, a retry that never takes an open chat back to
checking, the flag the ordering cases asked for, and the test harness the other
view models use.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 13:19:51 +02:00
Kgothatso NgakoandClaude Opus 5 db268a53d3 docs: plan a profile preview before a chat, starting from what the button is allowed to do
The "Direct message via npub" option goes straight from a pasted string to an
MLS room. The plan puts a screen between them, keyed by public key because that
is all the dialog has, and keeps the room's creation where it is: four decisions
-- where the preview lives, what "found" means when a row can be a placeholder,
that the button hands over rather than creates, and that it waits for the key
package so the twenty-second dead end is answered before the press -- and four
commit-sized phases with the tests for each.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 12:49:41 +02:00
Kgothatso NgakoandClaude Opus 5 29027f2b92 docs: record what the multiple-profiles plan built, and the places it chose differently
Phases 1–8 are implemented, in order, one commit each; Phase 9 is the
rollout and stays as written. The plan's phases are kept as the reasoning,
and the table at the top says where the build chose differently: the
repair reads before it writes and hands the listing its result; one
WalletAttached outcome with two ways in; the node stop and the relay scope
injected for their tests; the default save that cannot crash; a
ProfilesViewModel over flows; a NewProfileWriter and a route flag where the
plan expected the create screen's existing writer to serve; the colliding
id on the outcome rather than on the enum; the DAO's own requests left
unowned because they fetch public kinds; the inbox reopened by re-indexing
each wrap in its own transaction rather than by lifting the unseal branch
out; and the round trip's A made through the create view model with a
never-started PhoenixBusiness for the switch to stop.

Three things found on the way and in no phase are recorded beside the
table: the library's unsynchronised global-preferences cache, reached by
two threads at once for the first time, now behind JvmGlobalPrefs on the
jvm target; the same cache's consequence for tests that construct the
sovereign view model; and the gift wrap seal's link to its wrap, a foreign
key that existed and was never written until the inbox sweep asked the
question it answers.

The README's row and reading order say the note is built, and name
switchToIdentity where they named switchToWallet.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 01:57:05 +02:00
Kgothatso NgakoandClaude Opus 5 a5ff264164 docs: plan several profiles on a device, starting from what a profile is
The switch already exists as a state transition -- switchToWallet,
resetToSelector, a null identity sent to startup with popUpTo(0), every
collector a child of collectLatest -- and is reachable from nowhere: Landing
shows only when the device holds no identity, so a second can never be
added, and the profile tab's "change account" is a pending route. This plan
builds the two entrances and fixes what the transition gets wrong.

Before either, it changes one rule the two sign-in plans share. A seed's
nostr key is not written to the credentials file; it is derived from the
words at listing and from the running node at activation, and the writers
keep one key out of two files. That puts the wallet where the profile should
be: the list is a merge of two files with opposite ideas of what a row is,
and the node has to run for a profile to know its own key. Phase 1 makes a
profile a credential and a seed a wallet attached to one -- the credential
written when the seed is, repaired into the file for every seed already on
the device, merge inverted to list credentials and attach seeds by pubkey,
the identity's key read from the credential with the node's as a
cross-check, and a second refusal on forget for a key a seed derives. The
node still starts for a profile with a wallet attached, for the channel
watcher rather than for the key; making it lazy is now one branch and is
named as its own decision.

The other decision is that a switch is a restart of the signed-in graph,
not a swap under it: every route carries the key it was pushed for. That
settles the switcher as a pushed screen behind one tap, the previous node
stopped, the last-used profile as the one that opens on launch, and a
profile added from inside switched to.

Nine phases: the credential; the switch, with the relay observer that never
cancelled its predecessor and the node that kept running; the startup
precedence, which put "show me the list" above "open this one" and never
saved a default; the switcher, showing the nostr profile rather than
"Default name" and labelling a row with a wallet attached; the add rows,
where a profile created from inside is a bare key and not a second wallet,
and "end this" -- which wipes every profile's database -- goes; an owner on
the two fetch queues and an inbox sweep on activation, because a gift wrap
fetched under the other profile's key is stored and never opened again; the
exits; a round trip; rollout, with the one downgrade that lists a seed's
profile twice.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 00:09:02 +02:00
Kgothatso NgakoandClaude Opus 5 f866b9d17b docs: record what the npub sign-in plan built, and the twelve places it chose differently
Phase 8 of docs/npub-sign-in.md is the rollout, which is process rather than
code; what is left to write down is what the seven phases before it turned
out to be. The header moves from "Not built" to "Built", with the table the
other phased plans keep: what the plan said against what the implementation
did, twelve rows, each a decision worth reading before touching the code
it describes -- the startup branch one phase early because the sealed type
asked for it, the migration's three outcomes, the DAO guard that was
belt-and-braces on paper and load-bearing for two phases, a null identity
answering "can sign" with true because nobody is not read-only, and a
round trip run against a real database with a contrast case so that
"nothing was signed" is known to be a claim the harness can refute.

One finding that belongs to no phase is recorded with it: a compose test
looking for a floating action button's label has to search the unmerged
tree, or its "does not exist" is vacuously true. And the one rollout fact
that is not process: the library commit is on claude/nostr-credentials at
01962f3, bumped in by Phase 2, and like the nsec plan's before it has to
be pushed and tagged before any of this leaves the machine.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-12 20:52:53 +02:00
Kgothatso NgakoandClaude Opus 5 78807fe956 docs: plan npub sign-in, starting from what a read-only identity is for
The nsec plan's out-of-scope note said a read-only mode is a product, not a
branch. This plan takes that at its word: what a public key can see here is
thin -- a profile card, its follows as search results, no feed, no rooms,
since every room is MLS or a gift wrap to the key that was not pasted -- so
the first section decides what such an identity is for before anything is
designed. It is a preview: the app as your own profile, before you paste a
secret into it. That one word settles the Messages tab (an empty state that
offers the upgrade), pasting the nsec of a read-only key (an upgrade in
place under the same id, not "already on this device"), sign out (real for
this kind only), and not-found (try again or a different key, never set one
up).

Eight phases: a nullable key on Identity, with the note that the compiler
will be silent about it; a plaintext list beside the two key files, app-side
through the public getDatadir and AtomicFileWrite so nothing needs a JitPack
tag; the third startup branch, a read-only KeyPair built in one place, and
only the two pumps that read; the sign-in screen, where hex stays a secret
because an x coordinate is almost always also a valid scalar; a LocalCanSign
capability and an inventory of every write entrance one tap from the three
tabs; two exits; two round trips; rollout.

Two traps found on the way are recorded where they bite: quartz's
KeyPair(privKey = null) generates a fresh key rather than meaning "no key",
and decryptGiftWrapSeal forwards exactly that; and signInToProfile plants a
second kind 0 on a second call, which nothing reached until the upgrade
path.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-12 15:21:13 +02:00
Kgothatso NgakoandClaude Opus 5 2326839e48 docs: record what the nsec sign-in plan built, and the twelve places it chose differently
Phase 8 of docs/nsec-sign-in.md is the rollout, which is process rather than
code; what is left to write down is what the seven phases before it actually
turned out to be. The header moves from "Not built" to "Built", with the
table the other phased plans keep: what the plan said against what the
implementation did, twelve rows, each one a decision worth reading before
touching the code it describes -- the not-found exit as a screen state rather
than a route, input problems on the field rather than through ErrorState, the
kind 0 written over the placeholder rather than beside it, forget taking the
account rows too.

Two findings that belong to no phase are recorded with it: the android source
set's file-name collision that put the failure classifier in its own file,
and DataStoreManager's process-global preferences cache, which only a test
reusing a key across directories could have found. And the one rollout fact
that matters: the library commit is on claude/nostr-key-store in the
submodule, bumped in by the Phase 3 commit, and has to be pushed and tagged
with this branch.

The docs index now says the plan has been built and points at the table.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-12 12:30:45 +02:00
Kgothatso NgakoandClaude Opus 5 daae63d5c5 docs: plan nsec sign-in, and the ten call sites that make it small
An nsec is a BIP32 leaf of the seed at m/44'/1237'/0'/0/0, and the
derivation runs one way, so an nsec can never have a wallet behind it. What
makes the feature tractable anyway is that nothing in the app reads the
wallet except the nostr key -- ten sites, all the same expression -- so the
plan puts an Identity in front of the wallet and gives an nsec an identity
with no node behind it.

Eight phases: the identity type; a sibling key file in the library under the
one keystore alias Android will accept; listing and starting both kinds; the
sign-in screen for a phrase or an nsec, deduped on pubkey rather than wallet
id; indexer relays and a not-found exit for the sync that today asks one
relay and never finishes; recovery for a key with no phrase; tests; rollout.

Two findings along the way are recorded as pre-existing rather than new:
loadNostrProfile(startupRoute) keys off a Profile row a placeholder account
does not have and answers Landing while observeProfile answers the sync
screen, and the library's LocalKeyManager.nostrPublicKey() returns the
33-byte compressed key, not a nostr key.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-12 11:28:15 +02:00
Kgothatso NgakoandClaude Opus 5 e6aee79282 refactor: rebrand curated to curare, and the one name that stays on disk
`com.it.curated.{compose,android,desktop}` becomes `to.curare.{compose,android,desktop}`
across 700 file moves, and the app is Curare everywhere a user can read it. This is the
third rename this tree has been through, so the question was never how to spell the new
name but which of the 5074 occurrences of the old one were the brand. `3bc8be53` and
`d26cf6c7` had already sorted most of that, which is why this rename is mechanical where
those two were not -- and why the interesting part is the one place it is still not.

**`APP_DIR_NAME` keeps the old brand, and is the only thing that does.** It names the
desktop app's data directory -- `~/.local/share/curated`, `Application Support/curated`,
`%LOCALAPPDATA%\curated` -- and that single directory holds the Room database, the marmot
group state and the phoenix wallet. Renaming it does not move an install: it silently
starts an empty one beside it and leaves the old chat history and the old seed sitting
where nothing reads them any more. That is the same class of silent failure as the
`mantra/` hash prefixes, so it gets the same treatment -- it stays, with a comment saying
what renaming it would cost and that doing it properly is a first-run migration rather
than a rebrand. The desktop app is `curated` on disk and `Curare` everywhere a user can
read.

**The applicationId and the namespace are one value again, at `to.curare.android`.**
`3bc8be53` split them precisely so `com.it.curated` could land verbatim as the store
identity while the namespace kept `.android`; the decision here was to collapse them, so
both read `to.curare.android` and the manifest's relative `android:name=".CurareApplication"`
and `BuildConfig` resolve against the same package they always did. Debug is
`to.curare.android.debug` through the existing suffix. Worth stating plainly because
nothing in the build will: an applicationId *is* the app's identity to Play, so this is a
new listing rather than an update, and no installed Curated can update into it.

**`CuratedDatabase` -> `CurareDatabase` moves a directory, never a table.** The `@Database`
class name picks the schema *directory*, so
`composeApp/schemas/com.it.curated.compose.database.CuratedDatabase/` becomes
`to.curare.compose.database.CurareDatabase/` and its 19 files move unmodified. What makes
that safe has not changed since last time: the 12 entity classes are still `Mantra*` and
still declare no `tableName`, so Room still takes their class names as the SQLite table
names -- and this rename does not touch one of them. The check is that Room found all 19
schemas at the new directory and generated no twentieth.

**The generated resources package moved because `rootProject.name` did, again.** Compose
Resources derives it from the root project, so `Curated` -> `Curare` silently turns
`curated.composeapp.generated.resources` into `curare.composeapp.generated.resources` and
breaks 72 files' imports at once. Rewritten with them: the two string keys that carried
the brand -- `curated` and `sign_in_is_not_yet_available_while_curated_is` -- their three
call sites, and the eight `docs/scripts/` tools that hardcode the UI source path. The
audit reporting 582 `stringResource` sites and 178 `Icon(` call sites is the evidence
those tools still see the whole tree rather than an empty directory.

**A global rename edits the sentences that describe the previous rename, and two of them
came out false.** `material-design-conformance.md` recorded that `TorchTheme` became
`CuratedTheme` and that `UserAgent` was moved off Torch "when the app was rebranded to
Curated" -- which a blanket sweep turns into a claim about a change that had not happened
when it was written. Same in `HomeScreen`, where "the brand before that one" meant Torch
while Curated was current, and quietly came to mean Curated while still naming
`TorchTheme`. Both now say what happened, in order. The three deliberate Torch records --
the finding at :278, that note, and `NavigationRoutingTest`'s fixture -- are intact.

**What was protected stays protected, and now says "two rebrands".** `TWEAK_TAG`,
`HOST_KEY_DERIVATION_TAG` and `wss://ephemeral.mantra.press` are byte-identical; only the
prose around them moved, because "survived the rebrand to Curated" had stopped being the
whole story. The tree's 1668 `Mantra` and 17 `Torch` tokens differ from the 1666 and 16
before this change only by the words those comments added.

**iOS is `Curare` / `to.curare.ios`**, with the three pbxproj product references renamed
to match, and is unverified here for the same reason as last time -- the build gates the
ios targets behind `isMacOsX` and this was built on linux. The bundle id takes a platform
segment rather than the bare base it translates from, so that it reads alongside an
android applicationId that now carries one: `to.curare.android` and `to.curare.ios`, and
`to.curare` left as the base neither store claims.

1226 tests pass -- 787 in `:composeApp:jvmTest`, 439 in `:composeApp:testDebugUnitTest` --
`:composeApp:compileDebugKotlinAndroid` is clean, and `m3Audit` meets every budget.

Three stale things are left alone, because none of them is branding.
`.gitea/workflows/material-design-conformance.yml` still triggers on `branches: [mantra]`,
which stopped naming a branch that exists somewhere before this change and is now two
brands out; the three `torch_*` string keys still carry a brand two renames old, though
every value they hold reads Curare; and `androidMain/res/values/strings.xml` has opened
`<resources>a` since well before this. Each wants its own change.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-10 13:33:47 +02:00
Kgothatso NgakoandClaude Opus 5 3bc8be535a refactor: rebrand mantra to curated, and the five names that stay
`press.mantra.{compose,android,desktop}` becomes `com.it.curated.{compose,android,desktop}`
across 680 file moves, and the app is Curated everywhere a user can see it. The work
that mattered was not the rename; it was deciding which of the 6355 occurrences of
"mantra" were the brand and which were something else wearing it.

**"Mantra" was two words in one tree, and only one of them is a brand.** 1369 of the
tokens belong to a domain model -- `MantraArtifact`, `MantraChapter`, `MantraChunk`,
`MantraDialect`, the seven `MantraTranslation*` types, their 15 DAOs and
`MantraRepository` -- and those 12 entity classes declare no `tableName`, so Room takes
the class name as the SQLite table name. A blanket `s/Mantra/Curated/` renames 12 live
tables, which is not a rename but a v20 with 12 `@RenameTable` specs and a regenerated
schema. Asked, the decision was brand surface only, so the domain keeps its name and the
database keeps every one of its tables. The count is the check: the same 1369 tokens are
present before and after, and Room found all 19 existing schemas at the moved directory
and generated no stray -- which is what proves no table moved.

**Renamed instead is the set of names that only ever meant "this app".** `CuratedApp`,
`CuratedApplication`, `CuratedGlobal`, `CuratedNavHost`, `CuratedNavigationSuite`,
`CuratedShapes`, `CuratedConverters`, `defaultCuratedDir`, and `CuratedDatabase` with its
constructor and test. `CuratedDatabase` is safe to rename precisely because the entity
classes were not: the `@Database` class name picks the schema *directory*
(`composeApp/schemas/com.it.curated.compose.database.CuratedDatabase/`), never a table,
so the 19 schema files move unmodified and the AutoMigration chain from 1 is untouched.

**`com.it.curated` is the applicationId; the namespace is `com.it.curated.android`.**
These were one value before, and splitting them is what lets the identifier land verbatim
as the app's store identity while `BuildConfig` and the manifest's relative
`android:name=".CuratedApplication"` keep resolving against the android source package.
The debug build is `com.it.curated.debug` via the existing suffix.

**The generated resources package moved because `rootProject.name` did.** Compose
Resources derives it from the root project, so `Mantra` -> `Curated` silently turns
`mantra.composeapp.generated.resources` into `curated.composeapp.generated.resources` and
breaks all 501 imports at once. They are rewritten, along with the two string keys that
carried the brand -- `mantra` -> `curated` and `sign_in_is_not_yet_available_while_mantra_is`
-- and their three call sites. The seven `docs/scripts/` tools and `m3-audit.sh` hardcode
the UI source path, so they moved too; the audit passing is the evidence they still see
the whole tree rather than an empty one.

**Two hash inputs and a hostname keep the old name, and now say why.**
`SharedKeyDerivation.TWEAK_TAG` and `ChillDkgRitualManager.HOST_KEY_DERIVATION_TAG` are
`"mantra/..."` strings that are fed to SHA256. Renaming them derives different keys from
the same seed: every room already created is orphaned, every completed ChillDKG becomes
unrecoverable, and a Curated client and a Mantra client sharing a group derive different
keys from one threshold key, so their partial signatures aggregate to something that
verifies against nothing. That is a protocol fork and belongs in a `v2` tag, not in a
rebrand. `Relays.ephemeral` stays on `wss://ephemeral.mantra.press` for the duller reason
that it is a relay that is running and is in the bootstrap, publish, finder and DM sets;
pointing it at a name nobody has stood up takes the app off the network. All three sites,
and `docs/shared-key-derivation.md`, now carry the reasoning, because the failure mode of
each is silent and the next reader will otherwise finish the job.

**iOS had never been moved off the brand before this one.** `Config.xcconfig` still read
`PRODUCT_NAME=Torch` with a bundle id of `ac.aux.compose.Aux` -- two brands ago -- so a
Curated repo built a Torch.app. It is now `Curated` / `com.it.curated`, with the pbxproj's
three product references renamed to match. This is wider than Mantra -> Curated and is
called out for that reason; it is also unverified here, because the build gates the ios
targets behind `isMacOsX` and this was built on linux. In the same spirit the last stale
`press.auxiliary.compose` file -- a KMP template test asserting `1 + 2 == 3` -- was
relocated rather than deleted, so the source tree has one root package instead of an
orphan under the vacated one.

**Three things still say the old names on purpose and are not oversights.**
`UserAgent.APP_NAME` and `CLIENT_NAME` still read `"Torch"`, which a previous rename left
deliberately: it goes on the wire to relay operators and is a network identity rather
than a content one. `.gitea/workflows/material-design-conformance.yml` triggers on
`branches: [mantra]` because that is still the branch's real name. And
`docs/material-design-conformance.md:278` records that the top bar once read `"Torch"`
while the window title read `"Mantra"` -- a finding about a past defect, which rewriting
would falsify.

1136 tests pass -- 733 in `:composeApp:jvmTest`, 403 in `:composeApp:testDebugUnitTest` --
`:composeApp:compileDebugKotlinAndroid` is clean, and `m3Audit` meets every budget.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-10 00:53:05 +02:00
Kgothatso NgakoandClaude Opus 5 0180425904 docs: record what the subgroups plan built, and the six places it chose differently
All nine phases are built, one commit each. The phases are kept as written --
they are the reasoning, and the code reads better against the argument it came
from than against a summary of itself -- with a table of where the building
disagreed with the plan.

Six worth reading. `openCeremony` was never built, because a wrapper over two
repository calls the picker already makes would be a third name for one act.
`MarmotGroupCreation` is reached through the repository rather than called from a
view model, because view models here talk to repositories and managers take the
database. The guards' tests are in jvmTest rather than pure, because every refusal
reads the database and a pure version would test less. Phase 4 added two tags
rather than one, the second fixing a bug older than subgroups -- every robust
group has been arriving nameless on every device but its creator's. `stateFrom`
needed a third reader and a wrapper, because "tag present and unreadable" looks
identical to "absent" through a parser, and "refused" has to be distinguishable
from "none claimed". And Phase 8's two capability refusals short-circuited the
tests already written, which is how it came out that the fixtures had never had a
parent that could sign.

Two the plan got right and worth keeping if this is ever rewritten: the
key-package check moved to the picker on review, before a line was built, and it
is the difference between a subgroup failing in a second and failing after three
ceremonies; and the founding-roster rule has a test whose job is to fail the day
somebody adds the comparison that looks obviously missing.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-08 23:52:19 +02:00
Kgothatso NgakoandClaude Opus 5 c1ce262f3e docs: plan subgroups, and phase the four ceremonies a group needs to make one
A group can make another group, and the child can prove where it came from. This
is the plan for that, in nine phases, written against the code at b50b1762 and
not yet built.

**A subgroup is an ordinary robust group plus one artefact.** Fresh ChillDKG key,
fresh room, fresh quorum, and a birth certificate -- the parent's signature over
the child's room id -- carried on the child's `GroupKeyState`. Deriving the child
at `m/9420/1/0` instead would cost no ceremony at all and was rejected: a derived
child is the parent wearing a different hat, administered by the parent's members
with the parent's quorum, when the whole point is that a different set of people
can act on their own. The certificate is a claim about lineage, never a
delegation of authority, and nothing here lets one group sign for the other.

**Four steps, in the only order they can happen.** The ceremony produces `K`, so
the child's id exists; the parent's quorum certifies that id; the child's quorum
signs a key state carrying the certificate; the coordinator creates the room. No
step is a policy choice -- each needs the one before it -- and the last is gated
on the key state for the same reason `createAdminGroup` already is.

**What the parent's admins actually sign is the argument that shaped the event.**
Taken literally the certificate is 32 opaque bytes produced by a ceremony most of
them were not in. So the content is exactly the new group id as specified, and
the tags carry the child's threshold key, the path and the admin set -- covered
by the same signature, since an id hashes over its tags -- which lets a signer's
device check `marmotGroupId(key, path) == content` before agreeing, and lets a
coordinator who lies about who is in the child do it in a field the parent's
signature covers.

**The whole certificate travels as JSON on the key state, not a bare signature.**
A signature plus a rule for rebuilding the event it covers is a rule that breaks
silently the first time the event's shape changes: a rebuild differing by one
byte hashes to an id whose signature fails, and is indistinguishable from a
forgery. A parent tag rides beside it as an index into the certificate rather
than a second source of truth -- Phase 3 drops any state carrying one without the
other, or the two disagreeing, so there is no state where the index is believed
and the certificate is not.

**The ceremony stays on gift wraps, and the reason is `mls-skipped-keys.md`.**
Holding all three steps in the parent's Marmot room is the better design and the
plan says so at length rather than dismissing it: the certificate already runs
there, and the key state and the ceremony move together or not at all, since both
`GroupKeyStateManager.propose` and `signingPath` tie a key state to the room its
ceremony ran in. The mechanical cost is three enumerable changes. The reason to
wait is that the skipped-keys note already lists `proposeRitual` as a reliable
trigger, and a DKG cannot finish without every participant -- so one message
dropped for good stalls it permanently, where FROST needs `t` of `n` and routes
around a lost nonce. Revisit when the quartz fix lands; the collision Phase 4
refuses disappears with it.

**Three admins in total, and the threshold is set before anything is published.**
Three is `ChatRoomType.MINIMUM_ROBUST_GROUP_SIZE` for the reason that constant
gives, and the coordinator counts because they hold a share by construction, so
the picker asks for two others. `t` has to be chosen on that same screen and
nowhere later: ChillDKG hashes it and the host keys into the session identity, so
it is fixed the moment the proposal goes out, and a group that disagrees about it
gets no key rather than a weak one.

The nine phases are ordered so the checkable parts come first and can ship dark:
the certificate and its verifier are pure, the schema is three nullable columns,
and nothing produces a certificate until the button in Phase 7 exists. Phase 6
extracts the 120 lines of Marmot room creation out of `DkgRitualViewModel` so
both flows share the rules that are already right there.

What it does not do is named rather than left to be found: no revocation, no
delegation, certificates are not chroniclable, one subgroup per admin set, and
every selected admin has to show up twice.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-08 22:53:31 +02:00
Kgothatso NgakoandClaude Opus 5 1c02b25a07 docs: measure the UI against the M3 foundations, and phase the work that follows
A plan, not a change: what m3.material.io/foundations asks for as of its May 2026
revision, what these 43 screens actually do, and eight phases ordered so that each
one makes the next mechanical rather than judgemental.

**The spec was read, not remembered.** m3.material.io is a client-rendered SPA --
WebFetch returns an empty `<main>` and the tab URLs 404 on direct navigation -- so
the numbers here came out of a real browser session clicking through the tab
controls. That mattered: the May 2026 revision renamed window size classes to
**breakpoints** and there are now five of them rather than three (compact / medium
/ expanded / large / extra-large, at 600 / 840 / 1200 / 1600dp), renamed responsive
design to adaptive design, and published the spacing system as tokens on an 8dp
scale where `space100 = 8dp`. Writing this from memory of older M3 would have
produced a plan against a vocabulary the current spec no longer uses.

**The palette is fine; the call sites are not.** Every `onX`-on-`X` pair in all six
declared schemes clears 4.5:1, the tightest being `onPrimaryContainer` on
`primaryContainer` at 4.61:1 light and 4.56:1 dark. So the generated scheme is not
the problem and this plan does not propose a repalette. What fails is colour
decided locally, seven pairings of it, and the worst is not visible to a reviewer:

    Card(colors = CardDefaults.cardColors(containerColor = primaryContainer)) {
        ListItem(colors = ListItemDefaults.colors(containerColor = Color.Transparent),

`cardColors(containerColor = ...)` does derive `contentColor = contentColorFor(...)`,
so `LocalContentColor` inside the card is correct. But `ListItem` does not read
`LocalContentColor` -- its headline comes from `ListTokens.ItemLabelTextColor`,
which is `onSurface` -- and the call site overrides only `containerColor`. In the
light scheme `onSurface` and `primaryContainer` are both `#1B1B1B`. That is
**1.00:1**, and it is applied exactly to `proposal.awaitsYou`, so the proposals
waiting on your signature are the ones rendered invisible. `HomeScreen`'s
`titleContentColor = primary` on `containerColor = primaryContainer` is the same
mistake at 1.22:1. Ratios were computed rather than eyeballed; the script is in the
Phase 0 deliverable.

**Twelve colour roles fall through to Material baseline lavender.** `Color.kt`
never assigns `primaryFixed`, `primaryFixedDim`, `onPrimaryFixed`,
`onPrimaryFixedVariant` or the secondary/tertiary equivalents, so
`lightColorScheme()` defaults them to `ColorLightTokens.PrimaryFixed` ->
`PaletteTokens.Primary90` -> `#EADDFF`. Nothing reads them today, which is why it
has never been noticed; the trap springs the first time an expressive component
does. Read out of the pinned `material3-desktop-1.10.0-alpha05-sources.jar` rather
than assumed.

**Four of the six declared schemes are unreachable.** The medium- and high-contrast
variants are written out in full in `Color.kt` -- 78 colour values -- wired into
`lightColorScheme`/`darkColorScheme` in `Theme.kt`, and then never selected:
`TorchTheme` chooses between `darkScheme` and `lightScheme` only. The work to
honour a platform contrast setting is already done and disconnected.

**10dp and 20dp are not the problem they look like.** They are the two dominant
spacing values (132 and 115 uses) and both are *on* the M3 scale, as `space125` and
`space250`. The plan says so rather than proposing a sweep that would change
nothing. What is wrong is that none of the 520 `.dp` literals records whether it is
padding, a gap or a margin -- the three categories the spec gives different rules
to -- so nothing can be adapted per breakpoint later. About 101 are off-scale
(50dp x 53, 15dp x 14, 5dp x 10 and so on), and `Modifier.height(50.dp)` appears 49
times as the same copied spacer above the same copied error message.

**Findings that were measured and then dropped.** `outlineVariant` reads 1.61:1
against surface and `secondaryContainer` 1.65:1, both of which look alarming and
neither of which is a defect: M3's own baseline sits in the same range, and the 3:1
rule the spec gives is for clustered interactive containers, not dividers or tonal
surfaces. `onSurface.copy(alpha = 0.38f)` is the specified disabled opacity and the
spec exempts disabled states from contrast entirely. Reporting these would have
padded the count and cost the reader trust in the rest.

**The rest of the audit, in counts.** 334 string literals in composables against 2
`stringResource` calls, with title case throughout ("Edit Profile", "New Chat") where
the style guide asks for sentence case. Zero `Snackbar` across 26 `Scaffold`s. 16
copies of `Text("Something went wrong")`, none of which offers a retry. 90 of 240
typography reads on `label*` roles, which are for component text, while `display*`
and `headline*` carry 9 uses between them across 43 screens. 33 bare
`Modifier.clickable` with no minimum target, two of them text-height. Two
`BoxWithConstraints` and no window-size handling at all, on a project with a desktop
target whose own entry point already says so in a comment.

**Eight phases, ordered by what each unblocks.** 0 baseline harness, 1 theme,
2 spacing tokens, 3 accessibility floor, 4 content, 5 states and feedback,
6 adaptive layout, 7 motion, 8 guard rails. Tokens come before the call sites that
consume them; the accessibility floor comes before the adaptive work that would
otherwise double the surface to fix; guard rails come last so they lock in real
state rather than aspiration. Phase 6 is the only one that cannot be done
mechanically and the only one marked not reversible alone.

**What it deliberately does not decide.** Whether the target is
`MaterialExpressiveTheme` or `MaterialTheme` -- the pinned material3 ships the full
expressive set and the code already opts into `ExperimentalMaterial3ExpressiveApi`
in 66 places, but it changes default component shapes and sizes app-wide, so it is a
product call and Phase 1 raises it rather than answering it. Also out of scope:
whether the monochrome palette is right, the per-component specs, iOS (which only
builds on a mac, and whose HIG asks 44dp where M3 asks 48dp), and the three package
namespaces the UI currently lives across.

No code changes. `docs/README.md` gains the row and the closing paragraph's note on
how this one relates to the others.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-07 23:44:32 +02:00
Kgothatso NgakoandClaude Opus 5 ea11e8b233 refactor: call it a chronicle, and keep "archive" for what a user does to a chat
Archiving a chat is an ordinary thing a user will want to do to a conversation,
and it is not this. This is the group's signed record, handed to a member who
joined after the work was done so their room stops being empty. Two unrelated
meanings of one word in one app is a bug waiting to be written, and
`ChatRoom.archiveRequestedAt` is exactly where they would have met: a column on
the chat row, named for the thing that is not the chat.

So the whole feature is Chronicle now -- `press.mantra.compose.nostr.chronicle`,
`ChronicleEvent` (30327), `ChronicleRequestEvent` (30328), the three tags,
`ChronicleManager`, `docs/member-chronicle.md`. The kind numbers do not move;
only the words do.

**The wire tags move too**, `archiveId` -> `chronicleId` and `archivePage` ->
`chroniclePage`, which is free exactly once. Both kinds are new and there is no
old build to stay compatible with -- the design note says so in as many words --
so the alternative was carrying the old spelling on the wire forever to save a
rename that costs nothing today. The recipient tag stays `p`; it was never ours.

**Schema v14, because two things had the old word written into stored data.**

`ChatRoom.archiveRequestedAt` becomes `chronicleRequestedAt`, renamed rather than
dropped and re-added: while it is set it is the only record that a device with an
empty room has already asked the group for its history, and a device that lost it
mid-flight would ask again on its next launch, and the one after that.

The three `ChatMessage.messageType` strings become their `chronicle*` spellings,
rewritten rather than left to a legacy constant the way `dkgApprovalNeeded` was.
These lines cannot be regenerated -- a chronicle is announced once, when it is
requested, sent and applied -- and an unrecognised type is not skipped by the
transcript. It renders as an ordinary chat bubble, so "Caught up on 12 items"
would come back attributed to a member as something they said.

`MIGRATION_13_14` does both, because Room can rename a column and cannot rewrite
rows in the same breath. `ALTER TABLE ... RENAME COLUMN` needs SQLite 3.25, which
`getRoomDatabase` guarantees by pinning `BundledSQLiteDriver`, and the column is
in no index, no foreign key, and there is not a view or trigger in the database
-- so nothing has to move with it. Five tests hold the two halves apart: the
value survives, the column keeps its position, a room that never asked still
reads as never having asked, the three types are rewritten, and every other type
is left alone.

**`isArchivable` is `isChroniclable`**, on the "recyclable" pattern, and it keeps
its job unchanged: the allowlist that stands between a replayed
`GroupKeyStateEvent` and the apply path.

No behaviour change beyond the rename. 797 tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-06 16:27:36 +02:00
Kgothatso NgakoandClaude Opus 5 3ee1676a04 docs: plan handing a new member the group's signed history
A member added after the work was done sees none of it, and nothing in the app
will ever show it to them. Two independent reasons, and the second is the one
that surprises people.

MLS gives no history: a Welcome carries the ratchet tree at the current epoch,
not the transcript, and `MarmotInboundManager` drops anything from an epoch it
holds no keys for. That is forward secrecy working rather than a gap to close.

But group-signed events never travel at all. `FrostSigningManager.complete` says
so in as many words -- a signed event authored by the threshold key cannot go out
as an inner event, because the outbound pipeline would re-author it as its sender
and strip the group's signature off -- so every device *derives* the finished
event from its own `FrostSigningItem` rows. A member who was not in the session
has no items, and no later message carries the event. So the second problem does
not follow from the first and is not fixed by fixing it: even a member who could
decrypt the whole back-transcript would still hold nothing an artifact, chapter
or chunk could be built from.

Which makes an archive not a convenience but the only path, and fixes the line
the design has to hold: **it carries what the group signed, never the chat.**
Restoring the chat would undo forward secrecy on purpose, and a signed event is
the only thing a new member can check for themselves.

**The property the whole plan rests on is already true.** A room's id *is* the
group's threshold key derived at the room's path -- `GroupKeyState.verifies` and
`FrostSigningManager.signingPath` hold that invariant from their own ends -- so
`isSignedByGroup`'s three checks collapse to `event.pubKey == chatRoomId`, an id
check and a signature verify. No key state row, no threshold key, no path, no
lookup. A member who can name the room can verify its signatures, which is
exactly the position a new member is in, and it means the sender of an archive
does not have to be trusted at all.

**Two guards the plan makes non-negotiable.** Nothing on the inbound nip30303
path verifies a signature today, and that is currently correct: rumors carry an
empty sig and are authenticated by the MLS frame, so nothing on the wire has ever
claimed group authorship. An archive is the first thing that does, so the verify
is the feature's entire security rather than hardening on top of it.

And verification turns "group-signed" into an admission ticket for the apply
path, which is a wider door than it looks: a `GroupKeyStateEvent` is group-signed
and would pass perfectly, so an archive could replay a genuine old one and
re-point what the room signs with. The archive therefore carries an allowlist of
document kinds, checked outbound and independently inbound -- the same shape, and
the same reasoning, as the cap on `k` in frost-batch-signing.md.

**Push and pull, in that order of appearance and the reverse order of
importance.** Pushing an archive after the Welcome is what the question asked
for, and on its own it fails the way marmot-membership.md describes: it is an
application message in the epoch the add created, so one that beats the Welcome
there is dropped rather than deferred, silently, while the inviter sees a
success. So the joiner asks instead -- a request is proof it has processed its
Welcome, and it covers the reinstall and the second device, which no invite-time
push can. The push stays as a latency optimisation, deliberately phased after the
thing that makes it safe.

Nine phases: the verifier, the events, assembling an archive, applying one and
the sweep that lets pages arrive out of order, the request, the push, UI, the
cross-device tests, and rollout. The sweep needs no new table -- the inbound path
already stores every inner event it decrypts, so it is the shape
`FrostSigningManager.replayStoredMessages` already has.

Also written down, because it is the first thing this will be reported as a bug
for: an archive lets a new member *read* everything and does not let them sign
anything. `proposeSigningBatch` wants a secret share and a place in the ceremony,
and a group that re-runs its ceremony derives a different room rather than
re-keying this one. Closing that needs share resharing, which is a great deal
more work than this and is the thing to build after it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-06 13:52:33 +02:00
Kgothatso NgakoandClaude Opus 5 1448ed5ad8 docs(frost): record batch signing as built, and what rollout needs
Phase 7 of docs/frost-batch-signing.md, which is the phase with no code in it.

Nothing needs a feature flag. k=1 is the entire behaviour of the app as shipped
-- no caller batches anything yet -- and at k=1 every message is byte-identical
to the app before Phase 1: encodeProposal returns the bare event object,
joinPayload of one value is that value, and every plural branch in the
transcript is only taken above one. The doc now tabulates that rather than
asserting it in prose, since it is the claim the whole rollout rests on.

The one rollout constraint stands: before a caller batches, the group has to be
on a build that understands array proposals. There is no negotiation for it and
adding one is not worth it -- an old device refuses an array proposal outright,
so the failure mode is a batch that never reaches threshold and is abandoned,
visible in the transcript and costing a retry.

Also records what is left, which is nothing in the protocol: deciding what to
batch is a product question, bounded only by "a batch is only as available as
its worst item" and "GroupKeyStateManager.propose must never batch".

The phases are kept as written rather than rewritten into a description of the
result -- the code reads better against the argument it came from -- with the
two places the implementation chose differently (itemIndex over index,
DROP COLUMN over a table rebuild) marked in their own sections.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-06 05:00:21 +02:00
Kgothatso NgakoandClaude Opus 5 dff41d417d feat(frost): move a signing session's per-event columns onto FrostSigningItem
Phase 1 of docs/frost-batch-signing.md, which is added here as the plan the
next phases follow. Schema only: a session still signs exactly one event, the
wire is byte-identical, and every existing test passes on the moved columns.

## What moved, and why it had to

A batch of k events is k independent FROST instances sharing a signer set, not
one signature over k messages. That is forced rather than chosen: a Schnorr
partial signature is `s = k + e·x` with `e = H(R‖P‖m)`, so two messages under
one nonce R give two equations in one unknown and the secret share falls out.

So the five columns that enter that equation -- unsignedEventJson, eventId,
nonceRandom, aggregatedNonce, signature -- move to a child table keyed
(sessionId, itemIndex). What stays on FrostSigningSession is everything outside
it: the ceremony, the threshold, the derivation path, the signer set, and the
one approval.

itemIndex is protocol rather than presentation -- nonces and partial signatures
are joined positionally against it -- so getItems() orders by it and nothing
re-sorts. Spelled itemIndex rather than index to keep hand-written queries free
of backticks.

No itemCount column. The count is a COUNT(*), for the same reason signerIds is
derived from the ceremony's participant order rather than stored: a
denormalised count is one more thing that can disagree with the rows.

## Migration 9 -> 10

Manual, not auto: Room can create the table and drop the columns but cannot
copy between them, and the copy is the whole point. A session in flight at
upgrade holds its nonce seed and the aggregate it is already signing against,
and neither can be regenerated -- losing either makes the next pass derive a
different nonce for the same message and publish a second partial signature
over it, which is the extraction case. Both are copied verbatim into item 0, so
an in-flight session resumes as though nothing happened.

Removing the columns uses ALTER TABLE DROP COLUMN rather than the usual
create-copy-drop-rename rebuild. FrostSignerMessage and FrostSigningItem both
reference FrostSigningSession(id) ON DELETE CASCADE, and DROP TABLE fires
cascades -- with foreign keys enforced the rebuild would delete every signer
message and every item just written. Whether it does depends on Room disabling
foreign keys around migrations, which is not worth depending on when
DROP COLUMN cannot go wrong. It needs SQLite 3.35 and unindexed,
unconstrained columns; these five qualify, and getRoomDatabase pins
BundledSQLiteDriver on every platform.

## Invariants established here for the phases that follow

- signerIds and every item's aggregatedNonce are one write-once unit, applied
  by applyAggregate() -- items first in one transaction, then the session, so
  "some items aggregated" is unreachable and signerIds != null stays the gate.
- Signatures likewise, via applySignatures(); isSigned() counts rows instead of
  reading a flag.
- complete() verifies every signature before applying any event, so a batch is
  all-or-nothing rather than half-filed.
- itemsOver() gives each item its own 32 bytes of seed. Independent seeds mean
  an off-by-one in index handling produces a session that fails to aggregate
  rather than one that signs two messages under a single nonce.

signedEvent() and isAwaitingApproval() now take the item(s) rather than the
session, which propagates to the repository, the view model and the screen.
advance() reads items.first() and Phase 2 turns that into a loop.

## Tests

- FrostSigningSessionDaoJvmTest: index ordering, single-item read, upsert
  replacing rather than accumulating, signed-item counting, cascade delete.
- FrostSigningItemMigrationJvmTest (new): the backfill against a real v9
  database, asserting the seed and aggregate values survive -- not merely that
  a row appeared -- plus the exact column lists Room will check at open time.
- 338 jvmTest and 217 testDebugUnitTest pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-06 04:33:46 +02:00
Kgothatso NgakoandClaude Opus 5 af81933ab4 Merge branch 'mantra' into claude/room-db-testing-setup-b053cd
Brings the branch up to date with the 40 commits mantra gained while the
jvm target was being built, so that merging the other way is a
fast-forward.

One conflict, in docs/README.md, where both sides added rows to the index
table. Kept both, and gave the jvm-target note a clause in the closing
prose since it is the one document there that is not about the protocol.

One thing the auto-merge could not have caught. `9250991` added
NostrEventDao.getMarmotGroupNostrEventsByChatRoomId as a blocking query,
which android accepts and which Room refuses to generate for any other
target -- so the merged tree failed :composeApp:compileKotlinJvm with the
same "Only suspend functions are allowed in DAOs declared in source sets
targeting non-Android platforms" that phase 4 dealt with 58 times. Made
suspend; its only caller, NostrDao.reindexMarmotGroupEvents, was already
suspend, so again no cascade.

That is now a standing cost of this branch rather than a one-off: any DAO
method added on mantra while this is outstanding will break the jvm build
on merge. It is a one-word fix each time, and the compiler names the line.

Verified on the merged tree: :composeApp:compileKotlinJvm and
:composeApp:compileDebugKotlinAndroid green,
:composeApp:testDebugUnitTest 208 passing, :composeApp:jvmTest 214
passing -- both test tasks re-run from scratch rather than taken from the
cache.

The jvm figure is larger than the android one because jvmTest inherits
commonTest, so declaring the target quietly gained the whole shared suite
a second execution environment. That is worth knowing independently of
whether desktop ever ships: the same tests now run on the host, without an
emulator.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-06 02:01:35 +02:00
Kgothatso NgakoandClaude Opus 5 5abc37e463 docs: scope the jvm target, and separate it from testing the daos
Two questions arrived together -- whether Room's own testing guidance
applies to this project, and what desktop support would cost -- and they
turned out to have opposite answers. Both are now in docs/jvm-target.md,
phased, with the blocking work separated from the mechanical work.

**The expensive part is already done.** The four-deep native chain --
secp256k1 -> bitcoin-kmp -> lightning-kmp -> lightning-kmp-app -- already
builds for JVM, on every android build we do. The comment at
composeApp/build.gradle.kts:50 records the mechanism without drawing the
conclusion: lightning-kmp-core publishes no android variant, so our
android target resolves it to the *jvm* one, which pulls
secp256k1-kmp-jni-jvm desktop natives, which is exactly why the build has
to name the android artifact by hand. Read the other way round, every JVM
artifact in the chain is already compiled from source by the composite
build. A jvm target adds no cinterop, no C compilation and no new native
constraints. That was the part worth being afraid of, and it is finished.

**The blocker is one level down, and smaller than it looks.**
lightning-kmp-app/library declares 25 expects and implements them across
35 androidMain files. Its jvmMain holds exactly one: fibiprops.jvm.kt, the
Kotlin multiplatform library template's Fibonacci boilerplate, satisfying
two of the 25 -- both of them the template's own. So 23 actuals are
missing, which is why jvm() is commented out there
(library/build.gradle.kts:18), which is why it is commented out here
(composeApp/build.gradle.kts:46). Mantra cannot declare the target until
the fork does.

Six phases, ordered by that dependency. 0 build config; 1 the fourteen
mechanical phoenix actuals; 2 the three SQLDelight JDBC drivers and
NetworkMonitor; 3 key storage; 4 mantra's own sixteen expects; 5 the
desktop entry point. 1-3 are independent and parallelisable, 4 is where
the compiler finally checks the whole thing. Roughly a week to a
launchable build.

**Phase 3 has no day estimate, deliberately.** keyStoreEncryption /
keyStoreDecryption and their two graceful* wrappers delegate on android to
KeystoreHelper.kt -- 116 lines against AndroidKeyStore, StrongBox
attempted first and fallen back from, key material never leaving hardware.
Desktop JVM has no equivalent, so this is a decision rather than a port,
and the doc gives the three real options against what each actually
protects. A fixed-key JCEKS file is named there as a liability rather than
a stopgap: this is wallet seed material, and it lands on top of the
plaintext-key finding already open against this codebase. Recommended
sequencing is a passphrase-derived KEK with the desktop build marked
unsuitable for real funds, so phases 4 and 5 can proceed without the
security question being quietly treated as answered.

Two inherited mistakes are called out rather than carried forward. The old
Aux jvmMain put the database in java.io.tmpdir behind a TODO -- the doc
says not to inherit that in either phase that touches it. And
schedulePlatformLogic goes through WorkManager on android with no desktop
counterpart, so the doc asks for an explicit choice between a no-op and an
in-process coroutine, written down.

**The DAO answer is an appendix, because it is the opposite answer.** None
of the above is needed to test the DAOs, and burying that would have been
misleading. room3-runtime-android:3.0.1 already exposes the no-Context
inMemoryDatabaseBuilder(Function0<T>) overload, and
MantraDatabaseConstructor already supplies what it needs, so Room's
recommended host-machine form compiles in commonTest and runs under
testDebugUnitTest today. The one trap is native and is the secp256k1
problem mirrored: sqlite-bundled-android ships only android-ABI .so under
jni/, so a local unit test's JVM cannot load it and BundledSQLiteDriver
fails at construction; sqlite-bundled-jvm on the androidUnitTest classpath
is the fix. Robolectric neither helps nor is needed -- it cannot load
android .so on the host either.

Everything structural here was checked against the artifacts rather than
recalled: the Room builder overloads by javap on room3-runtime-android,
the two sqlite-bundled native layouts by unzipping both, and the
availability of room3-runtime-jvm, room3-testing, quartz-jvm and the two
SQLDelight drivers by request against the repositories this build actually
resolves from. The absence of android.* and java.* imports in commonMain,
and of any NFC reference from it, was likewise grepped rather than
assumed.

**Not verified: anything that requires compiling.** No jvm target was
turned on, nothing was built, and the day estimates are estimates. Phase 4
is where dependency-substitution surprises would surface if there are any,
and it is precisely the phase nothing here exercises.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-06 00:38:11 +02:00
Kgothatso NgakoandClaude Opus 5 39eac61838 Merge branch 'mantra' into claude/long-running-chat-sync-8983dc
mantra had moved on ~30 commits, several of them in exactly this area — and it
turns out both branches independently found the same bug and drew the same
conclusion about the same filter.

**The overlap.** f38a5f1 fixed the three kind:1059 filters that named the wrong
pubkey, including the two `authors=[userPublicKey]` requests in NostrDao that
could never match a wrap signed by a throwaway key. This branch deleted those
same two blocks, inverting the same `if` to the `== null` case, for the same
reason. The code merged to the same shape; only the comments conflicted, and
they are combined.

**Nip17Filters wins, and the live subscription now defers to it.** ad3304a
extracted the inbox filter to one definition precisely because it had been wrong
in three call sites, with the no-`since` reasoning this branch arrived at
separately. Keeping a fourth copy inside LiveSubscriptionManager would recreate
the problem that commit exists to solve, so:

  - queueCatchUpSynchronization now calls Nip17Filters.inbox() instead of
    building an identical SynchronizationFilter with its own limit constant,
  - Nip17Filters gains liveInbox(), the same shape as a quartz Filter for a REQ
    rather than a SynchronizationFilter for the queue, and giftWrapFilter()
    defers to it.

Two types for one filter is not duplication worth removing — the queue stores
one and hashes it for computeId, a live subscription puts the other on the wire
— but they belong side by side, because drift here means one of them quietly
stops matching mail.

**ChatMessageListViewModel keeps this branch's resolution.** mantra had it
refresh our own inbox on open (Nip17Filters.inbox on our DM relays, purpose
"chat"); this branch removed that call entirely. Both were right when written,
and the merge is where the second becomes true: LiveSubscriptionManager holds
exactly that filter open on exactly those relays for the whole account and
reconciles it on every foreground, so opening a chat has nothing left to ask
for. The redundancy is now recorded in the comment where the branch used to be,
so it reads as superseded rather than dropped. Discovery — the kind-10050 lookup
for a participant we cannot yet address — is untouched, and the purpose is no
longer a conditional now that only one case reaches it.

The commonTest coroutines-test dependency arrived on both sides; the comment
gives both reasons.

Verified: 154 tests pass, both branches' suites included — Nip17FiltersTest and
the marmot direct-message suites alongside this branch's 46.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 23:58:40 +02:00
Kgothatso NgakoandClaude Opus 5 bcdfd2ec94 Merge branch 'mantra' into claude/marmot-direct-message-type-7a0473
Twenty-two commits had landed on mantra since this branch left it, several
of them in the same files. Merged this way round so mantra stayed untouched
until the result compiled and its tests passed.

The migration had to be renumbered, and this is the conflict that mattered.
mantra is at database version 7 and already has its own 5.json -- for
MarmotInnerEvent.payloadEventId, nothing to do with direct messages. This
branch had also written a 5.json, for a different schema. Resolved by
restoring mantra's 5.json untouched and moving the direct message columns
to an AutoMigration(7, 8) with a regenerated 8.json. Taking either 5.json
over the other would have left every device validating a migration chain
against a schema it was never built from; keeping version = 5 would have
made a v7 install refuse to open at all.

The regenerated 8.json is two ADD COLUMNs and nothing else, same as before.

fromGroupEventResult was restructured on mantra: the kind switch moved into
applyInnerEvent, and a SubmissionEvent envelope now wraps nip30303 payloads.
Took that structure and re-applied the direct message branch ahead of it
rather than inside it -- a gift wrap is not a nip30303 payload to apply, and
what happens to it depends only on whether this device's key opens it, so it
does not belong in a function about applying submissions.

The isUserMessage fix was re-applied to the eight call sites mantra's
version has, up from the six it had here.

ChatMessageListViewModel and ChatRoomMessagingScreen took mantra's versions
with the composer state, the two renderings and the reply action layered
back on.

docs/README.md keeps both new rows and mantra's closing note about the
skipped-keys document.

108 tests pass, up from 50 here and 83 on mantra.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 23:45:08 +02:00
Kgothatso NgakoandClaude Opus 5 d110737f9a fix: keep a room's MlsGroup alive so a late message can still be read
Two events published in the same second reliably lose one of them. The
receiver stores the kind:445 and produces nothing from it -- no inner
event, no chat line, no error anybody sees, because MarmotGroupEvent is
written before the message is decrypted and so survives while everything
downstream silently does not.

Observed as a FROST signing session that never started on the receiver:
proposeSigning publishes the proposal and then the proposer's own nonce,
the relay handed them back in the other order, and the proposal was
dropped. The nonce is still sitting there filed against a session that
will never exist. The same bug ate a dialect earlier, which then took out
the artifact referencing it via a foreign key.

MLS is specified to tolerate this. RFC 9420 says a receiver that gets
generation N+1 before N keeps the intermediate keys so the older message
can still be read, and quartz's SecretTree does exactly that, in a
private skippedKeys map. What it does not do is persist it:
exportSenderStates() returns the ratchet positions only, so saveState()
drops the cache. NostrDao rebuilt the group from stored state for every
inbound event, so the cache was empty every single time, and generation N
arriving after N+1 failed `require(generation >= applicationGeneration)`
and was swallowed. Terminal -- the key is derived from a ratchet that has
moved past it, and nothing asks the sender to resend.

This keeps the instance alive instead. MlsGroupCache holds one MlsGroup
per room, and the inbound path goes through it, so skippedKeys survives
from one message to the next. That covers the case that actually bites --
a burst arriving in one sync, decrypted one after another against the
same tree -- which is what every bursty flow needs: proposeRitual sends
two, addArtifact sends two, and addChapter sends one per paragraph plus
one, of which only the ones arriving in ascending generation order
survived.

Reuse is conditional on the stored state still being exactly what the
cache last wrote. Sending a message advances the sender ratchet and saves;
so does adding a member. When that happens the cache rebuilds rather than
carrying on from a group that has been overtaken -- which is what keeps
this from being worse than no cache at all: the fallback is always the old
behaviour, never a diverged ratchet.

One lock per room, not one overall, because the group is mutable and
decryption advances it: two events for the same room decrypted at once
would corrupt the tree, and a busy room should not hold up a quiet one.

**This is a mitigation, not the fix.** It does not survive a restart, and
it does not survive another writer, so a long enough reorder still loses
the message. The fix belongs in quartz -- carry skippedKeys through
saveState/restore -- and quartz is a mavenCentral binary, not a fork, so
it cannot be made here. docs/mls-skipped-keys.md has the analysis, the
patch, the migration constraint on the persisted state format, and the
three ways to actually land it.

Not verified end to end: the proposal that exposed this cannot be
recovered, since its generation is already past, so confirming the fix
needs a fresh burst.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 23:09:53 +02:00
Kgothatso NgakoandClaude Opus 5 635cef9311 docs: write down how a direct message travels, and what it costs
The reasoning behind this is not recoverable from the code, which is the
bar docs/README.md sets for having a document at all. Three things in
particular would otherwise have to be rediscovered by whoever changes this
next, and two of them are traps.

Why the wrap uses a throwaway key rather than the sender's own -- and what
that does not buy. It does not hide the sender from the group: MLS
authenticates every application message to a leaf, so the identity is
there regardless. What it costs is a carve-out in MIP-03's pubkey check
and the sender's ability to ever read their own messages back.

Why the check that carve-out removes is not a hole. The authorship claim
moves from the wrap's plaintext pubkey to the seal's verified signature,
bound to the MLS leaf that sent it -- strictly harder to forge than what
it replaced.

The one query that would broadcast one of these. What this builds is a
genuine, correctly signed NIP-59 gift wrap, indistinguishable from what
the NIP-17 path would be right to publish, and the only thing keeping it
off a relay is that it never becomes a GiftWrapPayload.

Written against what shipped rather than what was planned, so it records
two deviations. senderIdentity is resolved in NostrDao rather than added
to GroupEventResult.ApplicationMessage, because quartz is a binary
dependency here and the local checkout is a reference copy, not a build
input. And a failed validation drops the message and logs rather than
throwing, because the caller is inside storeNostrEvent's transaction.

The unbuilt parts are listed as absences rather than left implied: there
is no member picker, so a private message can only be a reply to one
somebody already sent, and nothing in the UI yet tells a user in words
that the group can see who they messaged.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 19:10:22 +02:00
Kgothatso NgakoandClaude Opus 5 c8c962e4f4 docs: inventory the unreferenced code in the sync and relay stack
Found while building the long-running sync. One item was orphaned by that
change; the rest was already dead and only became visible because the subsystem
was being read closely. Written down rather than deleted because several pieces
are one decision away from being wanted, and those decisions are not the sync
change's to make.

Every claim is "this identifier appears exactly once in composeApp/src, at its
own declaration", with the two things that method cannot see called out: Room
DAO methods are reached through generated code, and Compose entry points can be
invoked without a textual reference. The DAO cluster is flagged as the least
certain for exactly that reason.

Three findings are more than leftovers:

  - RelaysSocketManager.userRelays is a field nothing ever writes. The
    `userRelays` inside observeRelays is a different, shadowing local, so the
    single-argument publishEvent always takes its FALLBACK_RELAYS branch and the
    user's own relay list is never used for publishing. That is a bug wearing
    dead code's clothes, and the fix is to populate the field, not to delete it.

  - NostrPublisherRepository is entirely unreferenced, and it is the only
    consumer of CachingImportRepository.importEvents. RelayPool and
    RelaysSocketManager each take a cachingImportRepository parameter they store
    and never dereference, satisfied by NO_OP_CACHING_IMPORT_REPOSITORY — so the
    whole seam is a parameter passed from nowhere to nothing. Removing the
    publisher lets the interface and both parameters go with it.

  - sendAUTH is unused because NIP-42 is unimplemented, not because it is
    surplus. AuthMessage is parsed and dropped, so a relay answering CLOSED with
    auth-required is retried forever and can never succeed. Deleting sendAUTH
    means deciding against authenticated relays; that is worth doing on purpose
    or not at all. sendCOUNT and CountMessage are a similar matched pair — both
    go or neither, since a CountMessage cannot arrive if nothing sends a COUNT.

isRecommendedRelay on the two request entities is separated out as its own risk
class: never written, never read, but a Room column, so it wants a migration
rather than a delete.

Ends with an order to do it in, cheapest and least risky first.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 16:28:21 +02:00
Kgothatso NgakoandClaude Opus 5 385c58ba7e docs: rewrite the sync note as what exists rather than what to build
The design landed across the six commits before this one, so the note is now
describing code. Reorganised around that: the reasoning that made it worth
writing is unchanged, but "the shape to build" is now "how it holds together"
and points at the classes, and the numbered traps have become properties of the
thing rather than warnings about a thing that did not exist yet.

Three sections earn their place after the fact:

  - the two timestamp decisions, which are the ones most likely to be "cleaned
    up" by someone who has not read this: no `since` on kind 1059 because our
    own wraps are stamped up to two days in the past, and no watermark on 445
    even though it would be safe, because `limit` already bounds the burst.
  - the four ways a group id can appear, which is why the group filter is
    derived from the room list rather than wired at the join sites.
  - "Not done", which was previously implicit in a staging plan: connectivity
    changes, NIP-42 AUTH, the collector-per-socket router the design originally
    called for, and the fact that the DM relay set is one relay.

The "suggested order" section is gone; git log is a better record of it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 16:20:08 +02:00
Kgothatso NgakoandClaude Opus 5 178ddd0181 docs: write down how a long-running chat sync would work
Every chat sync today is a pull: a screen queues a request row, a pump drains
it, the relay answers, the subscription is closed. Nothing arrives between
pulls, so a message sent one second after EOSE waits for the next time someone
opens a screen.

This note works out what it takes to hold the two chat subscriptions open for
as long as the app is active — kind 1059 p-tagged to us, and kind 445 h-tagged
with every group we belong to — and, more usefully, what in the current
pipeline quietly assumes a subscription is short:

  - completeOnSubscriptionEnd finishes the flow at EOSE, which is what releases
    the slot and sends the CLOSE,
  - SUBSCRIPTION_TIMEOUT hard-kills anything still open at 120s,
  - subscriptionSlots is a Semaphore(4) shared with the backfill queue, so a
    permanent subscription is a permanently-held permit,
  - both saveNostrEvent overloads need a request row to attach provenance to
    and to flip to "processed",
  - and nothing in the app reconnects a dropped socket at all. That is
    invisible today only because every subscription is short and the next
    queued request re-opens the socket on its way out.

The design keeps the queue and its three pumps exactly as they are: live
subscriptions replace polling, not reconciliation. Negentropy stays the tool
for first login, the catch-up after a background gap, and "load older".

The group filter is derived from chatRepository.observeChatRoomListByPublicKey
rather than wired at each join site, because a group id can appear four ways
and only one of them (creating a group) is somewhere anyone would think to call
a subscribe function — being added arrives as a Welcome processed deep inside
NostrDao.storeNostrEvent. Observing the room list also closes the loop: a
Welcome lands on the live gift wrap subscription, a ChatRoom row is written,
the Flow re-emits, and the group filter widens without anyone opening a chat.

Two findings fell out of checking the details against our own code:

  - `since = now` on kind 1059 would silently drop messages. Gift wraps are
    stamped with TimeUtils.randomWithTwoDays(), so a wrap published now can
    carry a created_at two days in the past. Kind 445 uses TimeUtils.now() and
    can take a watermark — opposite treatment for the two kinds we care about.
  - the "sent-messages" filter (kinds=[1059], authors=[me]) cannot match
    anything, because gift wraps are signed with a fresh throwaway KeyPair().
    It is also unnecessary: createNip17ChatRoom puts the user in their own
    participant list, so we wrap a copy to ourselves and the account-wide
    #p=[me] subscription already picks it up.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 16:01:47 +02:00
Kgothatso NgakoandClaude Opus 5 b99cb8fcd5 docs: write down the shared-key subsystem and how Marmot membership fails
First docs in the repo -- README.md is still the stock KMP template. Three
documents plus an index, covering the parts whose behaviour is not recoverable by
reading the code: where the reasoning lives in a protocol, where a failure mode is
silent, or where a decision looked arbitrary and was not.

marmot-membership.md is the one that earns its place. Everything about adding a
member compiles, the invite reports success, and a member simply never appears --
and the reason is never in the invite code. It records that
inviteMemberToChatRoom hardcodes isOneMemberInitialGroupCreation = false and that
ChatRepository does not expose it, so every group invite takes the deferred-welcome
path including the first, when the group is still just its creator and the commit
has no audience at all. Then why that is silent rather than noisy:
MarmotInboundManager refuses future-epoch messages outright, on both wire formats,
with no queue and no replay, so a commit arriving before its recipient's welcome
is dropped and that member never advances. EPOCH_RETENTION_WINDOW retains past
epochs and does nothing for messages from ahead. Three options are set out with the
per-invite correctness table, including the honest limit that the recommended one
narrows the race without closing it.

shared-key-derivation.md argues why the paths are not BIP32 -- no chain code
exists, hardened derivation is impossible rather than unimplemented, and a FROST
tweak takes the scalar as input so the chain code leaves the problem entirely. It
records the x-only serialisation trap avoided by choosing the scalar directly, and
states the rule that must not be broken: never reconstruct a derived key in the
clear, because k = k' - t hands over the group key rather than one derived key.

shared-key-ceremony.md covers the seven kinds, the three approval gates and why
the coordinator's aggregations are deliberately not among them, faults as values
rather than exceptions, and the transcript's idempotency-by-construction. It also
writes down the invariant that produces no error when broken: pendingApproval must
mirror the gates in advance, or the screen offers an approval that does nothing --
or none while the ritual sits still.

Every factual claim was checked against the source rather than recalled, which
turned up one correction worth having: there are two future-epoch refusals, for
PrivateMessage and for Commit, so the drop covers both wire formats and not just
one.

Each document leads with the failure mode rather than the architecture, on the
grounds that a failure is what sends somebody to docs in the first place, and each
lists its known gaps -- including that none of this has run on a physical device.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-05 14:39:19 +02:00