Three places where the copy promised more than the handler delivers.
The hero's personal-baseline line rendered unconditionally while both its
siblings in the same meta row gate on their own data. It claims the subtitle
above it was drawn from the reader's own last 90 days, but with no briefing the
subtitle is fixed fallback copy computed from nothing — so an account with no
briefing at all was told its generic sentence rested on a personal baseline.
Gate it on the briefing, like the rest of the row.
The pulse sub-page promised a personalised Karvonen target band. Karvonen is
not computed anywhere: the band comes from CDC/NCHS population percentiles by
age and sex, and falls back to a fixed 60-100 for a reader with no age on file.
The copy now says that in all six locales, and the fallback's own source label
says it is a general adult range rather than a personal one, because that arm
is the same band for everyone. Karvonen itself stays unimplemented; whether to
build it is a product decision, not a copy fix.
The Coach's `allTime` window returns 365 days, identical to `lastYear`, and it
is the default. "year so far" was wrong too — a rolling 365 days is not
year-to-date. Both labels now say what the reads do. The two options remain
functionally identical and one of them should probably go, but removing a
default is a product decision.
A rollup band outlives the rows it was folded from. The read-swap falls back to
live SQL only when a band is EMPTY, so a bucket that is merely wrong is served
as current; `ensureUserRollupsFresh` repairs the DAY tier over the trailing 90
days only, and keys off `Measurement.updatedAt`, which a hard delete does not
bump. The WHOOP body sync hard-deletes a WEIGHT row with no invalidation, so a
MONTH bucket keeps reporting a value the record no longer holds.
The block was handing the model two irreconcilable accounts of the same record
and marking neither: the coarse tail said one thing, the all-time extremes read
live beside it said another, and `asOf` stamped the whole thing current because
it is derived from raw reading age and knows nothing about rollup age.
Reconcile them where they are assembled. A bucket mean is an average of real
rows, so it cannot fall outside those rows' own all-time extremes — when it
does, the band is describing readings the live read cannot see. That is an
invariant rather than a heuristic, so the disputed tail is dropped and
`asOf.coarseHistoryWithheld` records that it went. Narrating a ghost is worse
than narrating less.
`currentForTodayClaims` stays keyed to raw recency, which is what it means. The
surviving readings really are from today, and suppressing a true present-tense
statement to punish a stale history band would trade one dishonesty for another.
The bucket on disk is still wrong, and a wide event now says so. The deeper fix
is to make the tier notice — compare a bucket's `computedAt` against the type's
in-window MAX(`updatedAt`), or invalidate on the hard-delete path — which is a
larger change than this repair.
The numeric grounding guard keyed its activation off `loop.toolResults` alone.
A tool that finds nothing resolves to `{ present: false, reason: "none" }` with
no `available` payload, `runCoachToolLoop` drops that shape, and the no-tools
snapshot fallback is populated only in the non-tool branch — so the guard sat
out every turn where the record was empty, which is precisely the turn where a
fabricated figure is likeliest and does the most harm. Ask about a metric you
have never logged and an invented number was persisted and streamed verbatim,
with no withheld-figure marker to say it had never been checked.
`toolTrace` is the discriminant, because it records the pure misses the results
array drops. Tools ran and every one missed: grade the reply — an empty ledger
there is not missing information, it is the positive statement that nothing can
be cited, so `findUnverifiedCoachNumbersInLedger` takes an explicit opt-in to
grade against it rather than no-opping. The model answered without calling a
tool: still dormant, unchanged, because the tool-mode base prompt carries no
pre-computed figures and flagging a legitimately recalled one was the v1.32.1
regression. The inventory counts ride along as a widener so a plain count
restatement stays grounded.
Eliding the digits is not enough on an empty record. "Your sleep averaged […]
minutes, up from […]" still asserts a series, an average and a direction for a
metric with no readings; only the precision goes, the false claim stays. So a
turn that missed every tool and lost a figure to the guard is replaced with the
honest answer, in all six locales. Scoped so a reply whose figure reconciles
against a prior turn or an inventory count is left alone.
Three characterisation tests for grounding gaps found while verifying that
every figure the Coach states comes from the record.
The numeric grounding guard rewrites any figure the turn's Grounding Ledger
cannot account for, and it holds: removing the strip reddens the existing
cross-turn ledger test. What is not covered is its activation. On the tool
path the guard runs only when `loop.toolResults` is non-empty, and a tool that
found nothing resolves to `{ present: false, reason: "none" }` with no
`available` payload, which the loop drops. So the guard disarms itself exactly
when the record is empty: ask about a metric with no readings, and an invented
figure is persisted and streamed verbatim, with no withheld-figure marker to
tell the reader it was never checked. The no-tools snapshot fallback is
populated only in the non-tool-mode branch, so the tool path has nothing to
fall back to, and every cloud provider runs the tool path.
Separately, a rollup bucket outlives the rows it was folded from. The read-swap
falls back to live SQL only on an empty band, `ensureUserRollupsFresh` repairs
the DAY tier over the trailing 90 days only, and it keys off
`Measurement.updatedAt`, which a hard delete does not bump. The integration
test reproduces the hard `deleteMany` in the WHOOP body sync against a
MONTH-band bucket: the deleted value keeps reaching the Coach's coarse tail
while the all-time extremes beside it report only the surviving value, and the
block is stamped current because `asOf` is derived from raw reading age and
knows nothing about rollup age.
All three assert current behaviour and carry a docblock saying so. Closing
either gap will redden them, which is the point.
The document defined bearerAuth and cookieAuth and referenced neither,
so the schemes were decoration: a client generated from the contract
wires no credential and reads the whole surface as open.
The default now offers the two as alternatives, and the ten operations
reachable before a credential exists opt out with an empty array. The
opt-out list is derived from what each handler accepts, not from the
proxy's public-path list — /api/ingest/medication is on that list and
requires a Bearer token, so reading it as the answer would publish an
authenticated route as open. Writing the list from the proxy is exactly
the mistake the guard caught: /api/auth/oidc/callback went on it and is
not published at all.
Two groups created environment.ts independently; one carried all six
paths, the other only the overview read, so the fuller side stands and
the barrel keeps one import.
Two component-id collisions, both from the same cause: two paths that
are two doors onto one action, documented independently.
DELETE /api/devices/{id} and DELETE /api/auth/me/devices/{id} both call
revokeDeviceCascade and answer with the same shape. Each grew its own
DeviceRevokeResponse, which the emitter refuses outright. The response
moves to shared.ts so neither module owns it and the next door onto the
same call finds it; the two envelopes keep separate names because an
envelope names an operation.
The settings module took both halves by hand rather than by keep-both:
one side added nine paths, the other three, with no overlap but a shared
import block and a shared reminder-thresholds entry that a mechanical
union duplicated.
Four surfaces answered a bad request by quietly substituting a good one. Each
fix comes with the test that catches it, and each test was confirmed red
against the old behaviour before the fix went back in.
`PATCH /api/user/ai-provider` inspected its body key by key with `typeof`
guards, so a numeric `model` or a boolean `baseUrl` failed every guard,
produced no entry in the update, and the write went ahead without it — a 200
that says "saved" for a field that was discarded. A body of nothing but such
keys came back as "No valid fields" naming none of them. The real schema now
lives in `src/lib/validations/ai-provider.ts`, the route runs one `safeParse`,
and a mismatch is the multi-issue 422 every sibling answers with. The accepted
field set is unchanged, and the object is deliberately not strict: an unknown
key has always been ignored here and still is. The one thing the schema has to
carry that no other does is that this body holds plaintext credentials, so no
rule may interpolate the value it rejected — noted where the next `.refine()`
would go.
`?slice=` on `/api/analytics` was compared against one literal and fell
through to the DEFAULT body on anything else, so `?slice=summary` answered 200
after running the heaviest chain on the surface. Invisible and expensive at
once. Closed enum, 422 on anything outside it.
`?metricType=` and `?limit=` on `/api/personal-records` were parsed
defensively, which read as robustness and behaved as silence — and in one
direction it was worse than silence: dropping an unrecognised `metricType`
WIDENS the read to every metric, so a typo returned more of the record than
was asked for. Both refuse now, and the 500 ceiling refuses rather than
clamps, because serving a different number of rows than were requested is the
same quiet substitution. No caller in the tree sends a value the new rules
reject.
`POST /api/ingest/medication` bucketed on the client address, so a household
running a chat bridge, a button and a home-automation rule — the normal
deployment for this route — put all three in one 60/min cell and let the
chattiest starve the rest. It already resolves a per-caller identity two lines
further down and threw it away; the bucket is that token now. A request that
resolves no token keeps the per-IP bucket it always had, through the shared
anonymous helper, so a flood is still capped and a broken proxy chain
collapses into the tight global cell rather than a shared "unknown" one. Its
429 goes through `apiError` and carries `meta` like every other refusal.
Two older tests pinned the defects by name — "drops an unknown metricType
silently" and "falls back to default on garbage ?limit" — and one pinned a
throw where the rest of the surface returns an envelope. All three are
rewritten with the reason beside them.
The generated YAML is deliberately not in this commit.
Twenty-five path entries across nine domains, all of them served and none
of them described. Same cause as the first batch: `openapi:check`
compares the registry against the YAML and never the route tree against
the registry, so a route registered nowhere drifts with every gate green.
Four groups where the contract had to say more than the shape does.
The full backup is the download that carries a whole record in plain
text, so what it contains is a safety question. It has two modes and this
route reaches only one: called with no options, it always emits the
PORTABLE form — decrypted columns, no tombstones, no row ids — while the
disaster-recovery form belongs to the worker and the admin path. Three
things it leaves out say so in the payload's own manifest: document
bytes, workout GPS and per-sample series, and the sensitive pair, where
the screener answers include the PHQ-9 self-harm item and a consent
receipt restored elsewhere would assert an agreement that operator never
obtained. The undeclared gap is named too, since roughly thirty models
the plan marks as backed up still have no reader or restore branch.
The four intake mutations differ in ways a caller has to know before
firing one. The by-id delete, the bulk delete and the purge all
tombstone, so the sync feed carries each removal; only the purge's
compliance rollups are dropped outright, and those recompute. The first
two refund the inventory stamp because not taking a dose puts the units
back; the purge deliberately does not, because refunding a whole history
would inflate the current container count instead. The bulk delete drops
unmatched ids silently, so a partial match succeeds with a lower count
than the ids sent.
The two provider-configuration surfaces are documented by what they do
NOT return. The chain answers provider types and boolean state, never a
credential. The per-medication ingest endpoint returns a raw token
exactly once, on the PUT that mints it, and a re-enable answers a null
token with `created: false` rather than re-issuing — the stored value is
an HMAC and there is no path back.
Insight generation costs money, so both gates are published with the
refusals they produce: the hourly per-user limit, and the day's token
ceiling, which is refused at reservation time before any provider is
contacted and carries its own code so a client can tell "too often" from
"the day's spend is gone". A 200 there does not mean a provider ran.
Two things the code said that were no longer true. The seeded-question
route's docblock declared it deliberately unregistered; it is registered
now, and a comment contradicting the registry is worse than none. And
the idempotency guard went red on `/api/insights/feedback`, whose
exemption said the path could not be published — which is exactly what
that leg exists to catch, so the entry is gone rather than the assertion.
`docs/api/openapi.yaml` is regenerated separately.
The drift check compares the registry to the emitted YAML. It has never
compared the routes on disk to the registry, so a route that was never
registered produced no drift and no failure: the document stayed
internally consistent while going quietly incomplete. Two hundred and
twenty-two paths reached that state.
The coverage guard walks src/app, reads the verbs each module exports,
and requires every verb to be either published or named with a reason.
The reasons are kinds rather than free text, because these fall into a
small number of genuinely different shapes and writing one sentence
seventy-three times would hide the entry whose sentence is different.
The security guard is the same idea one level down. The document defined
bearerAuth and cookieAuth and referenced neither, so a generated client
wired no credential and read the whole surface as open. The schemes
being present is what hid it. The document now declares them as
alternatives, and the guard checks the reference rather than the
declaration.
Both are red until the routes are documented, which is the point.
The eighth teardown endpoint, missed when the seven siblings were split across
the earlier batches. It has no per-user unsubscribe to make — WHOOP webhook
subscriptions are registered once per developer app rather than per category
the way Withings does it — and it leaves the BYO-key credentials in place so a
reconnect does not force a re-paste.
Like every other `POST /api/<provider>/disconnect`, a repeat call answers 404
rather than an idempotent 200, and the description says so. That completes the
tally: all seven disconnect endpoints refuse a repeat, all six credential
DELETEs accept one.
`POST` and `DELETE /api/notifications/web-push` both answered a flat
`apiError("Invalid data", 422)` and discarded the Zod issues. A non-HTTPS
endpoint, an endpoint aimed at an internal host, an over-long one and a missing
subscription key produced byte-identical bodies, so a browser whose
subscription was refused could not tell a fixable client bug from a policy this
instance will never satisfy. Both now answer the multi-issue envelope every
other body-parsing route in the tree uses.
Echoing the issues is safe here and the route says why, because it is the kind
of thing that gets re-derived wrongly later: the endpoint is a subscription's
routing secret and `keys` is its crypto material, but `sanitiseZodIssues` emits
`path`, `code` and `message` only — the rejected value lives in `issue.params`
and stays server-side — and every message this schema can produce is a fixed
string or a length/format default rather than an interpolation. A test pins
that none of the four secrets reaches the wire, so a future schema change that
adds an interpolating message fails here instead of leaking quietly.
The tests assert on content rather than status: a status-only check passes
against the flat error it replaces, which is how this survived. Two of them
compare whole bodies between two different refusals and require them to differ.
The two `/test` routes were checked in the same pass and need nothing. Neither
accepts a body — they take no parameters at all — so neither runs a Zod parse
to discard. Their 422s already carry `not_configured`, `vapid_not_configured`
and `rate_limited_self` as machine-readable codes.
Each of these made the server report something that was not so, and each now
has a test that fails without the fix.
The Withings credential DELETE wrote no audit row and left the ledger at
whatever state it last held. Fitbit, Google Health and WHOOP all park theirs,
and three of them say in their own comments that they do it for parity with a
list that never included Withings. A client reading `state` after the teardown
was told `connected` about a connection that no longer existed.
Oura's and Strava's credential DELETEs audited only when an access token
happened to be present, so removing a stored pair that had never completed
OAuth left no trail at all. Polar's twin has always done both unconditionally;
they match it now. The old Oura test asserted the skip, which is why the gap
survived a rewrite — that expectation is inverted here, with the reason next to
it.
The Withings and WHOOP manual syncs had no rate limiter, while `fullSync` walks
full history against a per-user provider budget. The Fitbit route's own comment
records that it was the outlier until it got one; these two were never brought
along. Both now carry the same 5/60s baseline and 1/hour full-sync buckets as
their siblings.
Every sync route parsed its body inside a catch commented "no body provided",
which was true for one of the three ways to reach it. An absent body still
means an incremental run — that is the documented shape and the web cards send
it. A body that is present and unparseable is now 400, and one that fails the
schema is 422: `{"fullSync": "true"}` used to be read as false and answered 200,
so a client that asked for full history and mistyped the value was told the run
it never requested had succeeded. Unknown keys are still ignored. The four
copies collapse into one helper.
The consolidated integration status swallowed a failed freshness query into an
empty array, which is exactly what a record with nothing in it also produces —
so the signal built to say "this metric has gone quiet" answered "nothing has
gone quiet" when it had failed to look. `metricFreshnessDegraded` separates the
two, as one flag rather than one per entry, because it is a single query for all
eight providers.
The contract follows the behaviour in the same commit.
`GET /api/medications/{id}/glp1` resolved the caller as themselves while
the POST beside it has resolved the record since v1.37.0. A manager
inside a shared record could therefore add a titration step to a
medication whose history the same path refused them: may write, may not
read. No reading of the sharing model produces that shape, and the
sibling reads on the same medication — `/inventory`, `/side-effects`,
`/cadence` — have all resolved the record at READ level for releases.
The read joins them at that level rather than at the POST's MANAGE,
because it is strictly narrower than the write already admitted. The
module was already on both reviewed allowlists, so nothing is newly
delegable; its reason line said the module was reached through the
MANAGE arm alone, which is no longer true and now says so.
The gap survived this long because a route that refuses every switched
caller is indistinguishable from a route nobody classified. Three legs
close that: a delegate scoped to medications reads the owner's titration
history and still cannot reach their own medication, a delegate scoped
elsewhere is refused, and a caller who never switched is unchanged. Both
delegate legs assert the refusal CODE rather than the status, because the
pre-fix answer was also a 403 — verified by reverting the handler, at
which point the first leg reads 403 for 200 and the second reads
`sharing.not_permitted` for `sharing.access.denied`.
Twenty-eight more verbs that exist, are called, and said nothing in the
contract: registration, password rotation, the profile patch, the SSO entry
point, passkey enrolment, the Codex device-auth flow, the device list and its
revocation cascade, four per-user preference scalars, the email channel and the
four channel self-tests, the record wipe, the privacy facts, the update check,
and the Web Push transport.
Four of them needed more care than the rest.
`GET /api/auth/oidc/login` is documented as the two-armed route it is. The
browser arm redirects to the provider; the native arm takes `client=native`
plus a mandatory S256 challenge and returns to a compile-time scheme with a
one-time code — or an MFA ticket when the account has a second factor. The two
PKCE exchanges in play are named apart, because confusing them is the way to
get this wrong.
`POST /api/auth/register` is the only genuinely anonymous surface in the set,
and its refusals are stated exactly: the SSO-only block, the 5-per-15-minutes
bucket that collapses rather than falls open, the invite that is consumed
whenever it is sent, the zero-user bootstrap that mints the first admin under a
lock, and the one 409 that covers a taken email and a taken username so it
cannot be used to enumerate accounts. `POST /api/auth/password` is a change
behind authentication, not a reset, so it is described as one.
`DELETE /api/settings/data` names what it takes and what it leaves: ninety-nine
models hard-deleted, the account's own personal columns reset, credentials and
interface preferences kept, one transaction, and a confirmation string that is
deliberately not the one account deletion wants.
Two collisions surfaced while wiring this up and are resolved rather than
worked around. `PUT /api/auth/profile` and `PATCH /api/user/profile` are two
doors onto one handler, so they now share one published request component
instead of two descriptions of one validator. Their RESPONSES differ, though —
one carries id and role, the other the display preferences, and neither is a
superset — so they are named apart and the older description no longer claims
to cover both.
The four Web Push and APNs surfaces land in their own route module rather than
beside the notification preferences: that module answers what gets sent and
whether it arrived, this one the plumbing a client sets up and then verifies.
Both self-tests report failure inside a 200, which the contract says plainly,
because a client reading only the status will tell someone their notifications
work when they do not.
The generated YAML is left out; it is regenerated once the remaining groups
land.
`openapi:check` compares the registry against the committed YAML, so a route
that was never registered produced no drift and no failure. These nineteen
had shipped without a contract: the health probe, the dashboard analytics
aggregate, the profile summary, personal records, the badge grid, the whole
environmental-context module, both onboarding writes, the Apple Health import
pair, the external medication ingest, the self-context adopt, the AI-provider
update and the AI connection probe.
Each entry is read off the route rather than inferred from a neighbour, so it
carries what a client cannot guess: that `/api/health` answers a bare object
and not the envelope, that `/api/analytics` records a score row and mints
pattern ids on a GET, that `/api/ai/test` reports a provider failure as 200
with `ok: false` because a 5xx comes back as somebody's HTML error page, that
the ingest route replays with 200 and creates with 201, and that a home
location re-stamps its effective-from instant on every write.
Three route-local request schemas move to `src/lib/validations/` so the
published shape and the runtime parse are one object rather than two copies —
a route module may only export handlers, which is why they could not be
imported where they were. The AI-provider PATCH has no schema to move: it
hand-rolls its parse field by field, so its entry says so and spells out that
a wrongly-typed key is skipped in silence.
The generated YAML is deliberately not in this commit.
Thirty-seven more paths the clients already call and the published spec did not
describe: the six OAuth providers' credential, status, disconnect, sync and
probe endpoints, the cross-provider surfaces under `/api/integrations/*`, the
managed record's configuration tree, and the native device-revoke path.
Where providers share a verb name they do not always share a contract, and the
descriptions say so instead of averaging it out. Three divergences are stated
per path. A disconnect refuses a repeat with 404 on six providers and answers an
idempotent 200 on two. `/api/<provider>/test` answers a bare probe result while
the same-named `/api/integrations/<provider>/test` adds the connection's last
sync, and the Google Health one accepts a body that replaces the response shape
outright. `GET /api/withings/status` refreshes and re-persists the OAuth token
while it reads; the other six status endpoints write nothing, and each now says
which it is.
Two further facts a client cannot infer. The Withings credential DELETE writes
no audit row and leaves the ledger at its last state, unlike every sibling, so
its `state` is not proof of a live connection afterwards. The WHOOP connect
ticket is a credential in a response body, returned once and stored only as a
hash.
The Google Health credential schema moves to the validations module; the managed
record's six strict patch schemas are exported from where they already lived, so
the contract publishes the objects the handler parses rather than a restatement.
Four defects the registry work surfaced, each with a check that fails without
the fix.
`GET /api/tokens` no longer consults the instance-wide API switch. That switch
gates the surfaces a token is FOR — external ingest, the MCP bridge — and not a
token's ability to authenticate: nothing on the `requireAuth` path reads it, so
a token minted before an operator flipped it stays live while it is off, and
refusing the list hid exactly the credentials their owner most needed to see,
their own signed-in phone among them. The revoke keeps the gate; the asymmetry
is stated in both route comments and pinned by the test, because widening it is
a decision about what the switch means rather than a fix to the read.
`PATCH /api/auth/me/unit-preference` reads its body through `safeJson` with a
1 KB cap, like the scalar beside it. It used a bare `request.json()` and was the
one endpoint in the family with no body-size limit at all — the new test proves
it by sending 2 KB and watching the old handler answer 200. Malformed JSON now
answers 400 like every sibling instead of 422; the only caller checks `res.ok`
and never the status, so nothing depends on the old code. The existing
malformed-JSON assertion was changed deliberately: it pinned the inconsistency,
not a contract.
`PATCH /api/auth/passkeys/{id}` returns the multi-issue envelope. It answered a
flat "Invalid request", so an empty name and a 200-character name produced
byte-identical bodies and the form had nothing to show beside the field.
`GET /api/user/thresholds` has a header that describes the handler. It promised
`{ defaults, overrides, effective }` and has always returned
`{ effective, overrides }`. The guard compares the two, so the comment cannot
drift again in either direction.
The two published operations whose contracts moved are updated with them.
Eleven routes the native client calls have never been in the registry, so
`docs/api/openapi.yaml` does not describe them. Nothing caught it:
`openapi:check` compares the registry against the YAML and never compares
the route tree against the registry, so a route registered nowhere drifts
with every gate green.
GET /api/analytics
GET,POST /api/cycle/symptoms/custom
GET /api/dashboard/summary
GET /api/environment
GET /api/export
GET,PUT /api/insights/settings
GET /api/insights/targets
GET,POST /api/medications/intake
GET,POST /api/medications/{id}/glp1
GET /api/mood/insights
GET /api/personal-records
Each contract is derived from the handler rather than from its comments:
the methods it exports, the parameters it reads with their real defaults
and bounds, the statuses it can actually produce, whether it resolves the
record fence, and which module gate it sits behind. Where the answer is
awkward it is written down rather than smoothed over — `/api/export`
answers a bare `data` key and not the standard envelope; `/api/analytics`
serves two unrelated shapes behind one query parameter and records the
health score it computes; `GET /api/medications/intake?scope=today`
projects rows before it reads them; the GLP-1 GET is not delegable while
the POST on the same path is; and `highlightInsight` on the dashboard
summary is pinned null by the handler and has never carried anything.
The intake aggregator's query and body schemas move from inside the
handler to `@/lib/validations/medication`, so the contract generates from
the same object the route parses with instead of a second copy of it. The
environmental overview gets its own route module because nothing owns that
surface. `docs/api/openapi.yaml` is regenerated separately.
`pnpm openapi:check` compares the registry against the committed YAML and
never the routes against the registry, so a route that was simply never
registered stayed invisible to the gate. Sixteen of them are called by the
native client and describe nothing: sign-out, the three per-user preference
scalars, the two pre-session discovery reads, passkey login options, the
passkey rename and delete, registration status, the consent latest-receipt
reader and its revoke, account deletion, the instance service switches, the
API-token list and revoke, the threshold overrides, the version read, and the
assistant feature matrix.
Each contract is derived from the handler rather than from its docblock, so
the published statuses are the ones the code can actually produce — the 413 on
a 1 KB body cap, the 400 the consent family answers where the rest of the API
answers 422, the 403 an instance-wide API switch produces on a token READ, the
last-admin and last-Guardian refusals on account deletion, and the fresh-factor
arm that a Bearer transport cannot clear at any scope.
Four request bodies moved to the validation modules the handlers now parse
with, so the published wire cannot drift from the runtime one: the three
`/api/auth/me` preference scalars into a new `user-prefs` module, and the
passkey rename beside the login schemas. Behaviour is unchanged.
The generated YAML is left out; it is regenerated once the remaining groups
land.
Sixteen routes the native client already calls were absent from the OpenAPI
registry, so the published spec said nothing about them. `openapi:check`
compares the registry against the YAML and never the routes against the
registry, which is why the gap survived.
Nightscout status / test / disconnect, the WHOOP credential and status reads,
and the Withings status / disconnect pair land in the integrations module. The
three per-channel configs under `/api/settings/*` join the settings module; the
notification preference and delivery-health surfaces get their own module, as do
the MCP connector connections and tokens.
These are credential-bearing surfaces, so the contract states what each response
actually carries. Every channel GET reduces its secret to a boolean, and the
descriptions say so, along with the consequence: a client cannot read the value
back, so an empty value on the way in preserves the stored one rather than
clearing it. `POST /api/mcp/tokens` is the one operation that returns a usable
Bearer, and its description says outright that this is the only copy.
Two inline request schemas move to the validations modules they belong in, so
the registry documents the same object the handler parses rather than a
restatement of it.
The register said a restored medication keeps today's schedule and loses
the record of when the dose or the timing moved. Measured, it costs more
than a record: the band minter reads each past day against the schedule
that was live THEN, so an account that came back without its eras had its
own compliance history rewritten against a plan it was never on.
The hard part was supersededByRevisionId, the pointer from a corrected
ARCHIVED era to the MANUAL row that replaced it. Three things about it
had to be settled.
It is a bare id column with no relation, so nothing in the database
refuses a value that points at nothing. That makes losing it quieter and
worse than a broken foreign key would be: every era consumer SKIPS a row
that carries the pointer, so a corrected era whose link came back null
goes live again beside its own correction, and the timeline shows one
window twice with past days minted against whichever the query read
first.
A portable restore mints fresh ids, so the pointer cannot travel as one.
It travels as a POSITION in the drug's own ordered era list. That list is
the only address space both ends share: the builder sorts it totally
(validFrom, then createdAt, then id) so two eras sharing an instant
cannot swap places between two exports of the same account, and the
pointer never crosses a drug, because every route that writes it scopes
the update to one medication. The same mechanism serves both purposes
rather than one carrying ids and the other positions.
Writing it needs two passes, because the row a link addresses is as often
written after the row addressing it as before. Pass one creates every era
with the link null; pass two patches the links once every position has an
id.
A hand-edited file that names a position out of range, or an era's own
position, gets the era and not the pointer, and the position is reported
as scheduleRevisionLink with the JSON path into the file. Refusing the
file would cost an operator every era of every drug over one bad integer,
and writing the link would point a correction at whatever row happened to
be there. A row that supersedes itself is treated the same way: every
consumer would skip it, so the era would vanish from the timeline while
still sitting in the table.
The round trip seeds three eras with one supersede among them and asserts
the shape of the chain, then asserts that exactly two are live. Three
rows with every link null satisfies any count and has thrown away the
only thing the pointer says.
The register said the drug comes back with no statement of what it was
for, which is exactly what a pinned efficacy target is: a row exists ONLY
where the person overrode the derived ATC / name resolution, so it is a
decision and never a computation. Nothing recomputes it after a restore
because there is nothing to recompute it from.
The lab arm travels as a biomarker NAME rather than an id, the way a lab
result's cross-reference already does. A portable restore mints a fresh
catalogue id, and `Biomarker` is unique on `(userId, name)`, so the name
is the only end of the reference that survives the trip.
The obstacle was ordering, not machinery: the remap map already existed,
but the drugs are rebuilt several hundred lines before the analytes are.
The targets are collected against the id each drug actually got and
written in a second pass once the biomarkers exist.
The alternative was hoisting the biomarker restore above the medications,
and it loses on two counts. `biomarkerByName` and `restoredBiomarkerIds`
are read by the lab-result loop immediately below the block, so moving it
would separate a map from its only other consumer and leave a later edit
free to re-order them back with nothing noticing. And the dependency
would read backwards: a pinned target is a leaf of the medication tree,
and letting a leaf dictate where the clinical record gets rebuilt inverts
what the file is about. The second pass states the constraint where it
applies instead of hiding it in a line order.
A named analyte the restore did not put back is dropped and reported as
medicationTarget rather than written. `biomarkerId` is a real foreign
key, so a dangling value would fail the constraint and cost the whole
account over one override. Dropping beats nulling here because the
override IS the reference: a row with neither arm set is what the
resolver already reads as "no override", so keeping it would restore
something that means nothing while still holding the primary slot.
The round trip asserts the restored target points at the biomarker the
restore itself wrote, resolved from the database rather than remembered
from the fixture. Counting would have passed on a row whose biomarkerId
came back null, and nothing else would have complained: the resolver
falls back to deriving a target and the page still renders.
The register said the originating device may still hold them. That holds
for about as long as the phone beside the watch does, and not at all once
a Withings connection is revoked: the signal id stops resolving with it.
So a self-hoster who exported and wiped had a file with no traces in it
and a restore that reported success.
Both ends live in src/lib/export/ecg-backup.ts, the arrangement the
reminders, visits and Coach sections already use.
Three things needed deciding.
The waveform follows the note contract: ciphertext verbatim in a
disaster-recovery payload, the micro-volt samples decrypted in a portable
one. There is no placeholder for a strip that will not decrypt the way
the Coach has one for a turn, because waveformEncrypted is a required
column. Such a recording is left out of a portable file with a warning on
the wide event, so the number of strips in the file is the number of
strips that can be shown.
measurementId is a REAL foreign key against the event row a strip was
filed with, unlike the Coach references next door. A value that will not
resolve does not dangle quietly here, it violates the constraint and
rolls the whole restore back over one cross-reference. A portable export
omits soft-deleted measurements, so that case is reachable rather than
theoretical. The reference is resolved against the measurements the
restore wrote, nulled when it cannot be, and reported as ecgReference.
The strips are wiped and rewritten by the section itself. The wipe at the
top of the restore never reached them: a recording hangs off the account,
not off a measurement, so deleting the measurements only nulled the
pointer and left the row behind.
The round trip asserts the row column by column rather than counting it.
A recording is three things at once and each is the whole value of the
row on its own: a strip with no rhythm verdict is a squiggle nobody has
read, a strip with no recordedAt belongs to no day, and a strip with no
samples is a claim that an ECG exists which nothing can draw.
Four models leave the coverage-pending register with both ends and a
round-trip proof: PersonalRecord, UserAchievement, EnvironmentContext and
EnvironmentTravelLocation. Two new section modules hold reader and restore
beside each other, the arrangement the visits, vaccinations, reminders and
Coach sections already use.
The register understated two of them. The bests are "recomputable in
principle", which is true and does not help: nothing recomputes them, and a
recomputation over a restored history would find different rows, because
sample series and GPS routes never travel, a portable file omits deleted
readings, and a record set in a reading that has since been corrected cannot
be found twice. The badge's unlock date is worse. The evaluator prefers a
persisted date over a derived one, and several of the dates it can still
derive come from counters no backup carries, so a dropped row does not relock
the badge, it re-earns it today.
PersonalRecord.sourceMeasurementId is the reference that needed care, and it
is the dangerous kind. The schema declares no relation for it; migration 0054
created the column as a real foreign key against measurements. Writing it
blind is not a dangling pointer, it is a constraint violation that rolls the
whole restore back and hands the operator an empty account. Measured, with the
resolution removed: "Foreign key constraint violated on the constraint:
personal_records_source_measurement_id_fkey", HTTP 500, nothing restored. So
the restore runs after the measurements, resolves the pointer against the rows
it actually wrote, nulls what it cannot resolve and reports it under a new
personalRecordReference skip kind.
An unlocked badge names a definition in the code catalogue, which drifts
between the release that writes a file and the one that reads it. The id is
carried and written back verbatim rather than judged, the same answer the
vaccination restore gives an antigen slug it cannot resolve.
The readings and the location periods travel as a pair because carrying the
readings alone is worse than losing them. The environment resolver reads the
periods live, so with them gone the nightly seven-day lookback re-resolves
every trip day to the home location and upserts over the row: an account would
come back with a fortnight abroad and then have it quietly rewritten as a
fortnight at home, days after a restore that reported success. Day keys stay
strings end to end, and fetchedAt is restored verbatim so a two-year-old
provisional reading does not claim it was confirmed today.
The round trip seeds bests that share a metric type and an instant and differ
only in the sport slot, two badges earned years apart (one of them under an id
this build no longer defines), and a trip day beside a home day, then reads
each back column by column. Counting alone passes every one of the failures
worth fearing here: slots dropped so a 5 km best is presented as a 10 km best,
unlock dates stamped at restore time, a trip day re-attributed to the home
city, a location period whose bounds no longer cover the reading it explains.
Each of those was applied on purpose and each failed on its own assertion.
Two schema comments claimed no read path selects a column that this backup now
selects; both say what reads it instead of staying wrong.
Reported from the iOS side, iOS #96. `GET /api/auth/native/login` answers
location: https://0.0.0.0:3000/auth/login?flow=native
behind a proxy that does not pass `Host` through. `NextResponse.redirect`
needs an absolute URL, the route built one with `new URL(path, req.url)`,
and `req.url` there is the address the Node process bound to.
That is the primary onboarding path. ASWebAuthenticationSession follows
the redirect, iOS refuses the restricted port, and the person lands on a
browser error page with no way forward. Distributed builds carry no
associated domains on purpose, so the browser handoff is also how
passkeys are meant to work; without it a new account is down to a
password and a confirmation code.
Two things hid it. Every error branch of that route builds the fixed
healthlog:// scheme with no host at all, so the route behaved correctly
on every failure and broke only when it was supposed to succeed. And its
own test built a request whose host matched, so the absolute URL it
produced looked right. The assertion there used toContain, which is
satisfied by the broken value just as happily as the correct one.
The same line was in six more places: the error redirects of the
Withings, WHOOP and Fitbit connect routes, which send a browser to
/settings/integrations. Behind the same proxy those land on the same
unreachable address, so all seven move to a relative Location.
A relative Location is valid per RFC 7231 and the client resolves it
against the address it actually called, which is the public one by
definition. The alternative is reading X-Forwarded-Host, and that would
put one more proxy setting between a self-hoster and a working sign-in.
Three checks, as the report asked for. The route test now drives a
request whose URL names the bind address, which is what a proxy hop
looks like from inside the handler. A structural test freezes the
pattern out of the four files and, separately, requires each of them to
still reach for the helper, because a file that simply deleted its
redirect would pass the first check and fail its users. The helper
refuses a protocol-relative value, which is a host in disguise.
Verified by breaking it: putting the old call back fails four checks,
the last of them reporting the reported value verbatim.
Refs iOS #96.
The scoped-sharing spec routes `/api/auth/me`, and the account switch it
drives ends in a full reload. The reload cancels whatever call is in
flight, Playwright marks the route handled, and the handler's `fulfill`
then lands on it with "Route is already handled!". With
`failOnFlakyTests` on CI that turns the whole shard red over a teardown
race rather than over anything the test is checking.
Same guard the record-session-fence spec and this file's own
accessibility sibling already use. The `fetch` is guarded too, because a
request cancelled mid-flight rejects there rather than at the fulfill,
and the earlier of the two is the one that decides which error the run
reports.
`DocumentConditionLink` and `ExtractedFact` were classified as backed up and
came back from no restore. The register said a restored vault is unsorted and
that the facts are re-derivable by re-running the extraction. The first is
right; the second understates it. What a re-run cannot produce is the REVIEW
DECISION on each fact, and every column holding one defaults to "nobody has
looked at this yet". `POST /api/documents/inbound/[id]/confirm` acts only on a
PENDING fact and commits it into the structured store, so a restore that let
those columns default would offer an already-committed reading for approval a
second time and write a second lab result behind it.
Both ends live in `src/lib/export/document-filing-backup.ts`, beside each other
like the Coach, reminder and vaccination sections. `status`, `needsReview`,
`confidence` and the commitment pair are carried and restored verbatim, and the
wire schema requires the first three rather than defaulting them.
Ordering: the section runs after the documents, the condition episodes, the lab
results and the medications. The first two are real foreign keys, unlike the
Coach's bare id columns, so an unresolvable value there does not cost one edge,
it aborts the transaction and the operator gets none of the account back. The
last two are what `committedRecordId` resolves against. The builder therefore
carries a filing or a fact only when both of its ends are carried, and the
restore still checks and drops what a hand-edited file leaves unresolvable.
`committedRecordId` is nulled together with its type, because a type with no id
claims a commitment the row cannot name.
The round trip seeds a second condition and files the page under it, so a
restore that kept the count and lost the pairing fails rather than handing back
a lab report filed under a head cold. A third case breaks both references in a
saved file to prove the drop path runs and still answers 200.
Three new skip catalogues arrive with it, and with them labels for the three
that had none: a dropped Coach attachment had been reaching the operator's
screen labelled "mood factor" since it was added. The label chain is exhaustive
now, so the next catalogue without one stops the build.
Two models come off the coverage register together, and neither is an
ordering or an id-remap problem. Both hang off the account and nothing
else. What kept them off was a question the code never wrote down, so
this change writes it down.
The screeners hold completed WHO-5, PHQ and GAD administrations, and for
PHQ-9 the encrypted blob contains item 9, the question about thoughts of
being better off dead. The application already draws a line around that
blob: it keeps a coarse item9Flagged boolean so the history view can
re-surface crisis resources without decrypting anything. The intent here
was to follow that line and carry the administration while leaving the
answers encrypted. The schema settles it more firmly. responsesEncrypted
is Bytes and NOT NULL, so there is no half-row to write and the choice
is the whole thing or none of it. A plaintext export in a downloads
folder is the wrong home for that answer, so the row rides in a
disaster-recovery file only.
The consent receipts are not sensitive in that way. Their problem is
that a consent is given to an operator. A receipt restored onto a
different instance asserts an agreement that operator never obtained,
with a signature timestamp to make it look settled. A disaster-recovery
file goes back to the same instance, so it travels there and not in a
portable one.
In both cases the portable file SAYS so in its own manifest rather than
reporting an account that has neither. The manifest is merged rather
than added beside the records one, because two keys of the same name
would shadow each other depending on spread order.
item9Flagged is restored from the file rather than recomputed.
Recomputing would mean decrypting every administration during a restore,
which is the handling this data is kept away from, and the stored flag
is what the history view has been reading all along.
Verified by breaking it: dropping revokedAt on the way back fails the
round trip with a withdrawn consent coming back as an active one.
Two entries sat side by side on the coverage register, described as "the
same shape of loss". They are not the same shape at all, and the
difference is worth writing down.
ReminderPhaseConfig hangs off a medication, one row per drug, nested
inside something that already travels. It needed nothing new and it
travels now. The round trip tunes all four thresholds away from their
defaults and puts one of them in PERCENT rather than MINUTES, because a
restore that quietly fell back to the schema defaults would hand back
four plausible numbers and change when the drug turns orange.
NotificationPreference cannot travel, and the register was wrong to
imply it was merely unbuilt. A preference addresses ONE notification
channel by a hard foreign key. The channel model is deliberately
excluded from every backup because it carries channel secrets, and the
restore deletes every channel it finds. A carried preference would point
at a row that cannot exist: the same constraint violation the custom
mood categories were causing, waiting for somebody to implement it in
good faith.
So it moves to the deliberate exclusions with a reason that names the
dependency instead of the symptom. Re-applying somebody's tuning to a
channel they add after a restore is a feature about channel types, not a
backup contract about rows, and it should be argued for on its own
terms rather than smuggled in as debt repayment.
An account that ever created its own mood category could not be restored
at all. Not partially, not with a gap: the restore violated
mood_tags_category_id_fkey and rolled the whole file back, so the
operator got a 500 naming a constraint and an empty account.
The coverage register described this as custom tags landing in a
grouping that no longer exists, which reads like a cosmetic loss.
Measured, it is the same shape the retired-catalogue-key work already
removed once: one row nobody thought about taking down everything else
with it.
MoodTag.categoryId is a real foreign key. Seeded categories belong to
the instance and are deliberately still not carried, but a category the
person made is theirs, cascades away with them, and has to travel. It
now does, and it is written before the tags that point at it.
The hidden set travels by key rather than id, because what people hide
is almost always a seeded tag whose id differs on every instance. A key
this instance no longer knows costs one hidden row and is reported: a
silent drop would un-hide a factor somebody had deliberately taken out
of their mood editor, and they would have no way to know why it came
back.
Found by experiment rather than by reading: seeding a user-owned
category into the round trip turned it red with the constraint name, and
the same test is green now.
A restored account reported no supply at all. Every pack was gone, and
with it every low-stock reminder, on a feature that became first-class
when medication moved to one stock per drug.
The two travel together because the register itself paired them: without
the events the count "cannot be rebuilt even in principle", and without
the packs the events are a list of deltas against nothing.
The count comes back exactly as recorded, not recalculated from the
ledger. It is the value the server had resolved and the person was
shown, and every reminder already sent was computed from it. The fixture
makes the difference visible on purpose: thirty in, eight consumed, and
a remaining count of twenty-one, because somebody dropped one without
recording it. A restore that "corrected" that to twenty-two would
disagree with the account's own history and look like a bug fix while
doing it.
Both decimal columns cross the wire as strings. Half tablets are a real
prescription and a JSON number would round them.
The note follows the contract the side effects and dose changes beside
it already use, and the packs nest inside their drug, so a portable
restore minting fresh ids needs no remap.
Verified by breaking it: recomputing the remaining count from the ledger
fails the round trip with twenty-one against twenty-two.
The transcript came back in the previous change; what the Coach keeps
BETWEEN threads did not. A restored account had every word of every
conversation and a Coach that has to be told its own history again — the
register said as much for all three: "a restore makes the account
re-tell its history", "keeps the talk and loses the commitment", "a
restored account silently stops sending them".
Tombstones ride only in a disaster-recovery file. All three soft-delete,
and a deleted fact is hidden from injection while kept for audit — a
distinction only the owning instance can act on. A portable file
carrying them would hand a person rows their own app refuses to show
them, and a restore that resurrected one would put a fact back into the
Coach's mouth that the person had deliberately taken out of it.
Neither reference here is a foreign key. sourceConversationId, on all
three, and relatedPlanId on the reminder are bare id columns, so nothing
in the database refuses a value pointing at nothing. That makes them
more dangerous than a real key rather than less: a dangling reference
costs no error, it just quietly stops meaning anything while the row
goes on looking complete. The restore resolves both against what it
actually wrote, nulls what it cannot resolve, and reports the count.
Nulling rather than dropping is the point. A live reminder must not be
lost because the plan it mentions was tombstoned out of a portable file.
The reminder is still the person's and it still fires; what it loses is
a pointer that was already going nowhere.
Order follows from that: threads, then plans, then reminders. Running
the memory before the transcript would cut every fact loose from the
conversation it came out of and still report success.
Verified by breaking it: making the conversation lookup always return
null fails the round trip on the fact's own source field.
A restored account got a Coach that had never spoken to it. The
conversations, every turn in them, and the record of which vault
documents a thread was grounded in were all classified as backed up and
travelled with nothing. The register's own entry for the messages called
them the largest volume of free text an account owns after its
documents.
The three travel together because two of them mean nothing alone: a turn
without its thread has nowhere to hang, and an attachment names a
pairing rather than a thing.
documentScoped is the field that makes this more than lost history. It
is a permanent fence — set when a conversation is created through the
fenced endpoint or gets its first document, and lowered by no code path
anywhere, not by detaching every document and not by deleting one. A
restore that let it default to false would hand the tool loop back to
precisely the conversations the fence exists to keep away from it, and
nothing in the interface would look wrong afterwards. So it is carried
verbatim, restored verbatim, and asserted per thread rather than in
aggregate: a restore that wrote false everywhere would still return both
threads and every turn and satisfy every count.
Prose follows the contract the rest of the payload already uses. A
disaster-recovery file carries the stored ciphertext; a portable file
carries the text, because a portable export exists to be readable by the
person who owns it, exactly as it already decrypts medication notes,
mood notes and document summaries. A single row under a key the instance
no longer holds is named rather than thrown, so one unreadable turn
cannot cost someone the rest of the transcript.
Ids ride in both modes, as the reminders do, so an attachment resolves
and the facts, plans and reminders still owed have something real to
point at. The restore runs after the documents for the same reason, and
an attachment naming a document the file does not carry is dropped and
reported instead of failing the whole file over one link.
Verified by breaking it: forcing documentScoped to false on the way back
fails the round trip on that field by name.
The restore refuses a portable backup, because a portable export carries
document metadata without the stored bytes and `contentEncrypted` is NOT
NULL. That refusal is deliberate and documented at the guard.
What nothing pinned is the half that makes it survivable: the check runs
above the transaction that wipes the account. The guard reads like input
validation, and input validation drifts toward wherever the value is
first used. Move it below the wipe and the operator loses everything to
a file the server was always going to reject — and the non-null
assertions further down are safe only because it stands where it does.
The test drives a real portable payload through the real route and reads
rows, not return values: the refusal happens, the account still has
every row it had, the cache invalidation never fired, and the audit row
names the document. Asserting the status alone would stay green through
exactly the edit this is here to catch, so the surviving rows are
asserted separately rather than inferred from it.
Verified by breaking it: disabling the guard turns the 422 into a 500,
which is the raw TypeError from decoding bytes that are not there.
Marc-André Bombeck