The harness landed in v1.30.35, after the slices that built this
release were already cut — so the reaction line's prompt module was
never registered, and its own docblock's warning against the
two-locale collapse was untested. Registering it surfaced a real gap:
the meaning-first instruction lived only in the system prompt's own
words, never in the user prompt in the family's shared phrasing, which
is where the rest of the surfaces carry it and where the tone contract
checks for it. The German body carried no such instruction at all.
Threads locale through the user-prompt builder so the fixed
instruction line matches the reader's own reply language, and wires
the per-user per-day opener-archetype rotation the milestone spec
called for and the first implementation had not connected.
The spine dispatches a generator-free enqueue; the worker resolves the
provider. Five gates stand in front of it in ascending cost order — both
modules on, ten-minute duration floor, a hard count of four a day in the
user's own timezone, an input hash over the deterministic evidence, and the
token ledger inside the shared status chokepoint. Each refusal returns a
status rather than throwing, because retrying against a ceiling that does not
move until midnight is an unbounded loop.
The detail route only reads: a workout with no row serves null and renders no
card, which is every workout that predates this and every re-synced one. An
undecryptable paragraph degrades to no card rather than failing the page.
Today gains two additive-nullable digest fields. `justIn` names the day's
newest arrival while it is still news — under three hours old, after which the
record is simply current rather than new. `reactionLine` carries a single
generated sentence about that arrival and replaces the hero's standing lead for
the rest of the local day, never appending a second paragraph, so the hero stays
one lead line tall and nothing below it shifts.
Nothing here polls. The open dashboard picks the change up on the cadence it
already runs — a state change, not a notification.
The generated sentence is bounded before it is written: the arrival spine's
unique row per kind per local day IS the throttle, the worker re-checks the
same claim before spending, and the reservation is reconciled on every exit
including each failure path. Without a provider, without consent, or against an
exhausted budget the row stays line-less and the surface falls back to the
deterministic lead — the chip and the provisional-to-final flip are unaffected.
The chip's time string is formatted client-side after mount; `justIn.at`
crosses the wire as an ISO instant so the server never renders a local
wall-clock time the browser would disagree with.
The evidence block is a numbers-only server projection: zone seconds, the
workout-window heart-rate shape, route climb and a same-sport 90-day own-history
median. Free-text metadata never enters it, and the sport string is narrowed
against the closed enum first, so the payload carries no text the user did not
generate through a measured field.
The prompt composes the shared assessment base and adds the session contract:
device-attributed, own-history-comparative, effort acknowledged where the
numbers show it, and no training prescriptions of any kind. Registered in the
tone harness beside the assessment cards.
The no-signal floors were the coldest text the product renders. They run when
there is nothing to ground against, and they opened on a bare order — "Measure
blood pressure at rest and under comparable conditions", "Evaluate mood trends
over several weeks" — so the one moment the surface has nothing to say was also
the one moment it told the reader what to do.
They now open on what the surface can honestly say: no assessment is available
right now. That is the meaning-first opening for this case, and it is the only
one available, because a floor with no signal may not imply a read. What
follows describes how the measure behaves rather than ordering the reader
around. The signal-grounded lines above them are unchanged.
The prompt version bump re-fingerprints the assessments cached under the old
biomarker prompt; the per-card input hash folds the constant in, so the
regeneration is one-time and paced by the nightly warm.
The per-biomarker assessment was the last surface in the family building its
prompt inline: a five-line scaffold that skipped the base system prompt, took
no opener hint, and instructed the model to "state the current value" — so the
one card where a bare number reads most like a verdict was value-first by
instruction while every sibling had moved to meaning-first.
It now composes exactly as the seven bespoke cards do. The new prompt module
carries a METRIC section that states the one thing this card needs and the
others do not: its snapshot has no signal block, so the read comes from the
pre-computed range placement plus the direction across draws.
Also point both archetype-driven cards at a registered budget. The generic
metric card and the biomarker card each carried temperature/maxTokens inline
at 0.45/1000, so the registry that documents every surface's ceiling was
silently wrong about the two widest assessment surfaces.
The single-marker no-key fallback replaces the multi-metric overview tip the
card used to render — text about watching several metrics at once, on a card
showing one lab value.
generateInsight() parsed against the strict aiInsightResponseSchema,
cross-checked that every recommendation's metricSource appeared in
citations[], and threw InsightSchemaError on a second failure. Its only
caller was runWithFallback(), which had no production caller — every
live surface runs runRawCompletionWithFallback. So none of it executed.
Wiring it into the live path was the alternative and it is not viable:
the live insightResultSchema has no citations[] field at all and treats
metricSource as optional, so the cross-check would have rejected
essentially every generation rather than tightening anything.
Removed: runWithFallback and its param/result types, generateInsight
and its private helpers, findUncitedRecommendations,
findRecommendationsMissingRationale, the never-thrown InsightSchemaError
and the isHardProviderFailure branch that excluded it, and
computeConfidence — a deterministic override that only ran inside the
dead wrapper.
Kept: buildRetryCorrectionMessage, which comprehensive-generate sends on
its own retry leg; the file it lived in is renamed to match what it now
holds. aiInsightResponseSchema stays as the shape that message
describes. Note the retry prompt asks for a stricter payload than the
live validator requires — that mismatch predates this and is left alone.
computeCitationCoverage is untouched and still live: it is a metric, not
a gate, and v1.30.24 wired it into the comprehensive post-parse step.
Chain-runner coverage the deleted tests carried (last-working cache,
empty-input guard) is ported onto runRawCompletionWithFallback, which
shares the machinery.
getBaseSystemPrompt appended the directive, then each metric module
appended its own section after it, so the directive was no longer the
last instruction the model reads - the one property that makes it
authoritative. The base now exposes its body, and every composing module
wraps the finished prompt so the directive ends it.
Also drops a duplicate import that survived the merge of the two prompt
branches.
Holds every sibling prompt module to one invariant: strip the directive
and the interpolated language name out of a French, Spanish, Italian or
Polish prompt and what remains is the English prompt, byte for byte. A
German instruction body leaking back into a non-German prompt, an
English body drifting away from what the four locales receive, and a
module appending a second directive on top of the one the base prompt
already carries all fail here.
Also drops the now-unused narrow locale alias from the two modules whose
prompt builders take the full locale.
The derived-score archetype and the Coach preference overrides branched
on the raw locale, so fr/es/it/pl composed German instruction text. Both
now route through `instructionLocale`. The score archetype rides on
`getBaseSystemPrompt` and therefore appends no directive of its own —
the base prompt already carries it.
The biomarker assessment is the one prompt in the family that does not
compose the base scaffold, so it appends the reader's directive itself;
the append collapses into the shared path once that prompt moves onto
the base scaffold.
The two prompt builders in each of those modules now take the full
`Locale` rather than the narrowed de/en alias, so the routing is right
the moment the pipeline stops collapsing the four locales on the way in.
Also targets the output-language contract test's regex at the compile
target instead of the dotAll flag, which does not compile.
German and English prompts are unchanged, byte for byte.
The per-metric system and user prompt builders branched
`locale === "en" ? EN : DE`, so a French, Spanish, Italian or Polish
reader composed the German instruction body — whose output clause asks
for German prose. Route them through `instructionLocale` instead, so
German is the only locale that takes the German body and everyone else
takes the English one plus the reader's own output-language directive
already carried by the base prompt.
The German and English prompts are unchanged, byte for byte.
The parity test joined raw fragments, which no longer matches now that
the English output clause carries a language token; it substitutes the
expected name from a literal map instead of from the module under test.
Adds the contract test: which body each locale composes, that its own
language is named, that its own translated directive is present and
last, that no German reaches a non-German prompt, and that no two
locales collapse to the same prompt. What it cannot show - whether the
prose reads well in the four languages - is stated in the file.
getBaseSystemPrompt collapsed every non-English locale to the German
instruction body, whose output clause asks for a German assessment. A
French, Spanish, Italian or Polish reader therefore received German
prose on any surface that reached this prompt.
Those four locales now compose the English instruction body with their
language named in the output clause, and the locale's own reply-language
directive - already carried in production by the native insights and
briefing prompts - appended last. German and English compose exactly as
before, byte for byte.
Adds the shared output-language helper the remaining prompt modules
adopt, replacing the scattered en-or-else-German binaries.
Fold the ECG silo into the insights narrative without crossing the
non-diagnostic line — the app reports the recording device's own verdict
and the existence/count of recordings, never an interpretation of the trace.
- Feature snapshot: a bounded, device-verdict-only ECG descriptor (recording
count, device-verdict distribution, latest recorded-days-ago + average HR)
flows into the nightly briefing prompt. The waveform never enters the
payload; the grounding gate admits the pre-computed counts so a
device-attributed sentence stays grounded. A prompt rule (EN + DE)
constrains any mention to the device's attribution and forbids reading
morphology, measuring intervals, or emitting a verdict of the app's own.
- Pulse cross-link: a device-attributed pointer from the resting-HR / pulse
page into the ECG viewer, un-mounting when no recordings exist.
- Daily rail: an ecg_new_recording item surfaces once when a recording landed
within the last day — calm, info-level, deep-linking the viewer. Gated on
the insights module; the DTO carries no waveform.
- i18n across all six locales; non-diagnostic disclaimer copy per language.
The per-metric insight assessments opened on the current value and read as
a data readout, while the overview texts lead with the day's verdict and
land the warm, motivating tone. Diagnosis (campaign insights-copy audit +
verification): identical warm persona/contract fragments, but the
per-metric USER prompt instructed number-first while the overview
instructed verdict-first, and the opener-hint rotation was wired into
briefing/scores/Coach but never the per-metric path (so the base prompt's
'do not open with the number' clause stayed dormant there).
Rewrote the seven per-metric assessment user prompts (generic archetype +
six bespoke: BP/weight/pulse/BMI/mood/compliance + overall) verdict-first
in EN+DE — meaning first, the number right after as support; grounding,
non-diagnostic framing, no-vendor and no-markdown all preserved (only
clause order changed). Wired openerArchetypeHint into metric-status and
the bespoke generators. Factored a shared openingShape contract into
shared-contracts so overview/per-metric/scores can't drift again, with a
coverage test. Lifted the three deterministic first-paint fallbacks to
verdict-led. Prompt-contract regression tests added. No new model calls,
token budget flat.
Thread the interpretation block into the generic per-metric assessment
path (the ~40 HealthKit metric cards). For a metric the knowledge base
covers, the prompt now carries the band table plus the server-computed
position of the current value, so the assessment names where the value
sits on the guideline scale and judges the trend by that position — the
visceral-fat 2.7 case now reads as "comfortably in the healthy band"
rather than reciting how many times it was logged.
The block is absent for uncovered metrics and for a sex-split band with
no known sex, so those cards fall back to exactly today's personal-relative
behaviour. Band edges carry no strip risk here — the generic assessment
path runs no post-generation numeric grounding gate.
Add two shared cross-surface contracts and compose them into the
assessment cards (base-system) and the comprehensive overview + daily
briefing (insight-generator):
- anti-recitation: a measurement count, a logging cadence, or page
mechanics is not an insight and may only appear when it carries a
consequence — stated as the consequence, not the count.
- interpretation-depth: say what a value MEANS, place it on its guideline
scale where one is attached, judge the trend by band position, and frame
the consequence without diagnosis; where no band exists, interpret
against the person's own baseline and say so honestly.
The cross-surface coverage guard now asserts both contracts reach the
surfaces that enforce them. The briefing's grounding and format machinery
is untouched.
Add a typed, hand-curated registry mapping metrics to guideline
interpretation bands derived strictly from the health knowledge base —
band edges, plain labels, valence, direction-of-good, and a source tag,
each entry citing the KB file + section it came from. A metric the KB
gives no numeric band for gets no entry, so its assessment stays
personal-relative rather than inventing a threshold.
The band position is computed server-side (`classifyBandPosition`, strict
upper-edge semantics + edge-hugging proximity), and `buildInterpretationBlock`
renders the table + placement into the assessment prompt so the model states
where a value sits and what the band means instead of reciting counts.
`interpretationBandEdges` exposes the edges for any numeric grounding gate.
Covers visceral fat, resting heart rate, blood oxygen, respiratory rate,
body temperature, sleep duration, waist (sex-split), waist-to-height,
pulse-wave velocity, and BMI. Tests pin every band edge against the
guideline value so a silent edit breaks a test.
The shared formatting contract asks every narrative surface for short
blank-line paragraphs, but two field-level instructions overrode it:
the assessment OUTPUT FORMAT pinned 'flowing prose (2-4 sentences)'
and the briefing paragraph asked for 'a connected 70-160 word read' —
so both consistently arrived as one unbroken block regardless of the
contract. Both now spell out 1-3 short paragraphs separated by a
blank line written as \n\n inside the JSON string.
The shared contract banned every structure beyond 1-3 paragraphs, so
longer replies read as unformatted blocks. It now asks for short
blank-line paragraphs (2-4 on longer replies), permits a plain "- "
list when a reply genuinely enumerates three or more parallel items,
and one bold key takeaway per reply. The Coach ground rule keeps
prose-first; a list is the exception, never the default shape.
Replace the fixed three-beat assessment skeleton with a menu of moves the
model orders freely, and reframe the daily briefing from per-signal
enumeration to a verdict-first read that weaves the few salient signals into
one associative story. Score cards now name both the helped and the hurt
contributors, translate the band into a forward read for the day, and close
on one grounded nudge; the deterministic fallback gains the same band-meaning
sentence so provider-less and demo accounts read as well.
Add a hash-seeded opener-archetype rotation (deterministic, testable) so
consecutive cards and days no longer open the same way, a day-rotated sampling
seed for the period narrative in place of the fixed reference seed, and a
shared paragraph formatting contract. Add a score-number grounding gate and
extend the briefing grounding allow-set to the glucose, labs, preventive-care
and workout aggregate blocks so real figures can be cited without opening the
gate. Gate the daily score warm on an input fingerprint so a quiet day skips
the model call.
Add a deterministic narrative layer the Coach narrates rather than invents:
- Briefing callbacks + forward-looks: the daily briefing recalls the prior
period and points ahead, anchored to a figure, one callback per surface.
- Tension Verdict: when the readiness composite's contributors disagree
(good sleep but a rising resting pulse), surface a brief honest internal
read on the health-score card. A real clinical red-flag is never reconciled
away into a calm verdict.
- Streaks + return-to-baseline: a pure detector over a DAY-bucket series
closes a worry once a metric returns inside the personal MAD band after a
genuine out-of-band run; a gap day breaks a streak.
- Coach prompt: add develop-discrepancy, roll-with-resistance, and an
anti-persuasion clause; ratchet PROMPT_VERSION.
The tension flag and streak event are computed server-side; the card lines
render resolved fields and the model narrates them.
The acute red-flag clause hardcoded its blood-pressure numbers as prose
literals and named no glucose or fever floor at all, so the Coach could
silently drift from the dashboard hero and the notification engine. Compose
the systolic/diastolic, glucose, and sustained-fever thresholds straight from
clinical-floors.ts and echo the previously-missing glucose and fever lines, so
every acute number the Coach states is bound to the one source of truth.
Add an outlook contract (gentle forecast, what-to-expect, anticipatory if-then
— all conditional, ranged, association-framed) composed beside the tone
contract on the Coach and the comprehensive briefing, so a reply can sharpen
expectations without a false promise. Tighten the tone contract against
over-validation (affirm the effort, not an unsafe choice; stay neutral when
there is nothing to praise) and route why/conflict questions through
get_correlations and its coincident-deviation flag before answering.
Coverage asserts the clause numbers equal the constants and the new contract
reaches both surfaces.
The retrospective narrative was the one model surface that omitted the
shared tone contract, so it read colder than the daily briefing beside it.
Compose the tone fragment into both prompts and warm the hand-written tone
line to name a genuine win when the period earns it, while keeping the
descriptive-never-causal and no-alarm guards intact. The cross-surface
coverage test now asserts the contract reaches this surface too.
The shared contracts cover chronic deferral (dose, diagnosis, drug-level
to a clinician) but had no acute branch. Add a closed-list crisis clause —
chest pain, syncope, sudden severe symptoms, hypertensive-crisis readings,
suicidal ideation — that points to prompt or emergency care without
diagnosing, and keep every other turn calm and non-alarmist. It rides the
single shared-contract source so it lands on the Coach, the per-metric
status cards, the comprehensive briefing, and the period narrative; the
coverage test asserts it on all four surfaces.
BP, pulse, weight, BMI, mood and medication-compliance still asked the
model to derive the recent-vs-baseline comparison from raw buckets — the
exact arithmetic small models do unreliably, and the pattern the
base-system prompt already advertises as precomputed.
Build the SIGNAL block (current / baseline / signed delta / normal-swing
verdict) with buildMetricSignal from each card's graded series and inject
it into the snapshot; BP carries two signals, compliance one per
medication. Each card's prompt section now leads from the signal block
and is told to state, not recompute. The server-computed compliance %
is unchanged.
One-shot prettier pass so the new format:check gate starts from a clean
baseline. Formatting only — no behaviour change. Generator-owned and
vendored artifacts stay untouched via .prettierignore.
Remove the unused output-key reverse map the batch path never read, and
have the budget table satisfy the AiBudget interface directly so the type
is enforced rather than orphaned.
Add generateStatusBatchForUser: it prepares the seven specialised status
snapshots once, sends one prompt for a { perMetric } envelope covering only
the metrics with data, and fans each summary back into the same per-metric
cache rows the standalone generators write. One JSON-correction retry guards
the parse; any omitted metric or a failed batch degrades to that card's own
single-call path, so a partial response never goes dark. The nightly warm and
the forced warm route through the batch in production while the injected
per-card seam keeps the cron tests intact. Adds the status-batch prompt
scaffold composed from the shared assessment contract.
The grounding, tone/banned-opener, GLP-1 dose-safety, metric-identifier
ban and forbidden-filler rules were independently encoded across the
comprehensive, status, Coach and narrative prompts and had drifted. Move
the canonical wording into `prompts/shared-contracts.ts` (locale-keyed
fragments) and compose each surface from it.
Where the surfaces had drifted, adopt the stricter wording: most notably
the GLP-1 dose-prescription contract — previously present only in the
comprehensive and Coach prompts — now reaches the status cards and the
period narrative too. A cross-surface coverage test asserts every
enforced fragment appears verbatim in every surface, in de and en.
Compute a present-focused "signals of the day" block in extractFeatures from
the in-memory measurement set (no extra query): the freshest reading vs the
user's own 7d/30d windows, the normal-swing verdict, an emerging slope, and a
recent anomaly per salient metric, ranked by clinical priority and capped at
five. Add a signalsOfDay array to the dailyBriefing schema feeding the
briefing alongside the existing keyFindings cap, and rewrite the briefing
prompt (EN + DE) to lead with NOW: a warmer, motivating read that surfaces the
emerging signal and gives one concrete, doable nudge while still drawing on the
tiered history.
The signals block lands inside the compacted features payload, so it also
seeds insightsSnapshotHash — a fresh daily signal now changes the fingerprint
and forces the daily 04:30 warm to regenerate instead of re-stamping a
days-old narrative. Bump PROMPT_VERSION to 4.29.0.
Add WHOOP-native DAY_STRAIN (0-21) to the Coach snapshot as a dedicated
dayStrain block, kept distinct from the computed STRAIN_SCORE (0-100) and
preferred native-over-computed, mirroring recovery resolution. Document the
glucose.clinical panel (TIR/GMI/eA1C/CV) plus the stillLearning and
isSpotEstimate gates in the system prompt, and the native dayStrain scale in
the derived paragraph. Extend the DE and FR/ES/IT/PL evidence-block unit lists
with mg/dL, mmol/L and min so a non-English coach can cite glucose and duration
values. Export GLUCOSE_PANEL_WINDOW_DAYS as the single source for the 30-day
clinical-panel window and import it across the analytics route, dashboard
snapshot, and Coach snapshot. Compute the derived metrics concurrently.
A typed apiFetch wrapper owns the envelope unwrap, the ok-check and the
error read for every client call — two hundred call sites across the
app migrate onto it and a lint rule keeps raw api fetches out for good.
The legacy timezone helpers fold into the tz module and test fixtures
drop personal names.
Warm the insight generator (briefing + key findings + insight prose) and
the Coach to the same premium-motivating advisor register already shipped
for the per-metric assessments. The intro and role framing drop the cold
clinical-summariser voice for a warm advisor who reads the user's data
against their own history, names earned wins, and frames an unfavourable
finding as one doable step.
Every safety and quality guard is preserved: no diagnosis, no
prescription, no platitudes, no bare number-echoing, honest about
unfavourable values, the banned positivity openers, the doctor-consult
call-to-action, and the JSON output schema + keyFindings tone enum stay
byte-identical. PROMPT_VERSION 4.27.0 -> 4.28.0 (also covers the Coach,
which imports it). Native FR/ES/IT/PL personas warmed to match.
The per-metric and wellness-score assessment prompts gain a premium-advisory
layer: name the earned win, build a little momentum, frame an unfavourable
finding as an opportunity, and close forward-looking. Every existing guard
stays in force — no platitudes, no bare number-echoing, autonomy-supporting,
never alarming, moralising or diagnostic, and the banned-opener and
forbidden-phrase lists are unchanged. The composite-score archetype warms to
a WHOOP/Oura-style recovery coach while staying descriptive, not clinical.
PROMPT_VERSION 5.0.0 -> 5.1.0 so the provenance can slice prose before/after
the tone edit. Cached accounts converge to the new voice within ~24h.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add an additive, non-breaking assessment field to GET /api/insights/derived:
{ text, source, updatedAt } | null, keyed to the requested metric id for the
five per-score ids (READINESS, SLEEP_SCORE, RECOVERY_SCORE, STRAIN_SCORE,
STRESS_SCORE). The field is null when the metric is not one of those or when
its status is not "ok". A deterministic text always fills it — composed from
the score's own contributors, so provider-less accounts and the demo are never
empty — and warmer provider prose overrides it once cached, warmed out of band
so the GET never blocks on a model round-trip. de/en, following the v1.12.12
deterministic-narrative model. The batch grid read returns assessment: null;
the detail sheets use the single-metric route.
Introduce a shared buildMetricSignal() helper: a token-lean, signal-rich
descriptor (current, own baseline, signed delta, spread, normal-swing verdict,
normal-range placement, contributors) computed server-side so the model states
the comparison instead of re-deriving it from raw buckets. Thread it into the
generic per-metric assessment snapshot and lead the base prompt from it. Add a
PROMPT_VERSION constant to the base system prompt so prompt changes are
traceable.
Regenerate docs/api/openapi.yaml.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The v1.12.1 diversity block (variety lead, data-strength line, steady-run
repetition signal, FDR-screened cross-metric relations) reached only the
generic archetype cards. The seven hand-tuned headline cards — the ones users
see most — ran at temperature 0.3 with only the previous-context block, so they
stayed prone to repetitive openers and forced filler steps. Thread the same
assessmentContextBlock through the specialized get*UserPrompt builders and
raise their temperature to 0.45 to match the archetypes.
Soften the unconditional "close with one step" across the specialized sections
and the activity / body-composition / sleep archetypes to the base prompt's
conditional form — close with a step only when the finding implies one,
otherwise affirm and name one thing to watch — so a steady metric no longer
manufactures a filler step. EN/DE parity preserved; guard computeSteadyRun
against an absent graded slice.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The insights ground-rules carried changelog tokens ("v1.4.16 phase B8", "W4d")
inside the prompt strings the model reads — token waste and inconsistent with
the project voice. Move the rule content to clean numbered rules; the markers
belong in code comments, not shipped prompt prose. Update the test pins to the
cleaned rule headers.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The per-metric assessment prompt (base-system) and the scope-hardened
insight system prompt (insight-generator) each carried EN and DE as two
parallel template literals kept in sync by hand — a standing drift risk
flagged in the AI-quality audit.
Model each prompt as one section skeleton: a single ordered list where
every section supplies its EN and DE fragment side by side, and a builder
joins the fragments with a blank line. The section order and prose are
unchanged; the DE calibration is preserved verbatim. A section edited in
one locale but not the other now shows up as an asymmetric edit in one
list instead of two literals silently diverging.
Pure refactor — every emitted prompt string is byte-identical for both
locales across every metric, surface and reference combination. A new
locale-skeleton-parity test pins the structure (each section carries both
fragments, composed prompt equals the joined fragments); the existing
grounding, filler-ban, forbidden-phrase, refusal and exact-phrasing guards
pass unchanged.
(cherry picked from commit 12cd718d6d602a48cafc14ad06d82282fdaacb27)
The per-metric assessment cards are tightly grounded but their form had
converged: every card opened the same way, forced a "one doable step"
even when nothing was actionable, and a metric that stayed steady for
weeks produced near-identical paragraphs. Squeeze more value from data
already computed server-side, without touching the grounding floor or any
safety contract.
- Surface FDR-surviving cross-metric correlations on the per-metric card.
The discovery engine already screens behaviour×outcome pairs with a real
Pearson + exact p-value + Benjamini-Hochberg FDR control; that
intelligence only reached the period narrative. A new read-only consumer
runs the same full-matrix discovery and filters to the pairs that
involve the current metric, then feeds the engine's own conservative,
descriptive, never-causal interpretation strings into the prompt as
grounded relations. Best-effort: a correlation hiccup can never block the
generation it only decorates.
- Streak-aware repetition. Derive a steady-run length from the graded
series (consecutive recent weeks within a band of the user's own
baseline) and tell the model when it has already reported "no material
change" N times running, so it acknowledges the continuity in one clause
and pivots to a different facet instead of restating the same level.
- Stop forcing a step; add controlled variety. The closing step is now
conditional — when nothing is genuinely actionable the model affirms and
names one thing to watch rather than manufacturing the filler that lets
platitudes back in. A deterministic per-render variety token (seeded from
user + metric + day, never Math.random / Date.now) rotates the opening
angle so consecutive cards and days don't read identically, and the
phrasing temperature lifts 0.3 → 0.45 while the facts stay pinned by the
snapshot and the forbidden-phrase guards.
- Surface the computed data strength (n + recency) into the prose prompt so
it hedges on thin data instead of guessing what "few" means, matching the
UI confidence badge.
- Add locale-matched few-shot examples (a grounded assessment + a labelled
banned-filler counter-example) to the assessment system prompt, which
most helps weaker local-provider models follow the contract.
The own-baseline grounding, computed-not-hallucinated stats, schema-
enforced output, filler-phrase ban, and correlation-as-association
discipline are all preserved and re-pinned by new guard tests.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
(cherry picked from commit c9d04cab1523e0e57a3f50c87c9f88684b01b26f)
Fold the deterministic short-horizon forecasting engine into the Coach
prompt as a compact, lowest-signal snapshot block. For each in-scope
metric the block carries direction + slope + the projected horizon end
with its widening prediction band, the last observed value, and the
fit's R² / server-computed confidence — read straight from the existing
projection engine, never recomputed. A metric below the fit/history/
staleness floor is omitted; the block is dropped entirely when none
qualify, and it sheds first under prompt-budget pressure.
Add a conditional-projection ground rule so the Coach narrates a
trajectory only when the block is present and only for the metric it
covers: direction and range, always conditional, never a certainty, a
risk score, a diagnosis, or a dated event; absent block means no
projection. The rule lands in the safety-contract matrix across all six
locales and inline in the EN + DE system prompts, with an overclaim
probe asserting a projection is never upgraded. Bump PROMPT_VERSION.
Extend the assessment-card system beyond the seven specialised scopes to
the ~30 HealthKit metric pages, without hand-writing a prompt per metric.
- Add a metric-metadata registry keyed by HealthKit id (display name,
unit, direction, optional normal range, archetype, scope), covering the
data-bearing HealthKit metrics. The seven specialised metrics keep
their own scopes and are excluded from the generic set; an unregistered
id gets no card.
- Add a generic data-driven generator that reuses the existing graded
snapshot, provider runner, stale-while-revalidate cache, and persist
path. One prompt template per archetype (physiological-vital,
activity-fitness, body-composition, mobility-gait,
environmental-exposure) plus a dedicated sleep template, parameterised
by metric metadata. Empty-data metrics return an insufficient marker
with no provider call. Normal ranges use age + sex where the profile
stores them; no ancestry/region ranges.
- Add GET /api/insights/metric-status?metric=<id>: apiHandler-wrapped,
Zod-validated against the closed registry id set (422 on unknown),
read-only with stale-while-revalidate. First insights route with
OpenAPI coverage; regenerate docs/api/openapi.yaml.
- Route the generic metric:<id> scope through the on-demand status queue
worker dispatch.
- Extend the nightly warm pass to the generic metric caches for a user's
data-bearing metrics only (one distinct read gates the set), keeping
de+en warm and never calling a provider for an empty metric.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
(cherry picked from commit 41453fbdfee60793c16269b9a9e5f04b6c3ea450)
The hard '~5 measurement points' floor did not fit mood (counted in
entries) or medication adherence (counted in scheduled doses). Re-word
the gate as 'a few measurement points/entries/events' and note that the
unit depends on the metric. Also fix the gating-field path: the recency
field lives at dataCoverage.newestMeasurementDaysAgo, not on summary.
The mood card noted persistent low phases but offered no gentle path to
support. When mood stays clearly low for several days, the model now
offers an autonomy-preserving pointer toward talking it through with a
trusted person or a professional — never alarming, never a diagnosis, no
emergency tone. The safety-contract matrix does not cover this pathway
and does not reach the per-metric assessment prompts, so it lives here.
For an infrequent schedule such as a weekly injection, the 7-day
compliance figure is dominated by a single dose: one miss swings the
percentage sharply and reads as a collapse it is not. Tell the model to
lean on compliance30 as the load-bearing value when schedulesPerDay is
low and treat compliance7 as a rough pointer only. Output stays {summary}.
The overall-status card summarises up to six metrics, but it inherited
the single-metric 30-60 word budget, which forced the model to either go
generic or overrun. Grant the section an explicit 3-5 sentence / 50-80
word budget and cap it at naming two metrics by name, with the rest
mentioned only when they deviate notably from their own mean.