Keeps the voice-core prompt rules and adds main's world-query and
vessels-vs-visuals rules; the tool catalog wraps the deixis-aware runner.
Also fixes two things the branch had before the merge:
- Live voice could not start: OpenAI rejects track_entity's top-level
anyOf. Model-facing tools now omit top-level combinators; the runner
and benchmark still require a query or a referent.
- The Contacts-centred analyst answer dropped the viewport coverage note.
A sweep for the same pattern found more answers presented as current and
complete: launches and satellite orbits served from a stale proxy cache, a
regional aircraft fallback read as worldwide, license plate cameras trimmed
or partly loaded, and a radio directory of popular stations read as every
station. The launch and satellite sources now report the proxies' stale
markers, and each answer says what it covers.
- Voice layer manifest (src/voice/layerManifest.js) generates the layer enums,
aliases and analyst field hints; a test fails when a catalog layer has no entry.
- analyst_query validates fields, operators, units, values and scopes, counts up
to 250,000 loaded records per layer, reports complete/partial/unanswered, tells
LAYER_OFF / NOT_READY / FEED_UNAVAILABLE apart from zero, refuses mismatched
follow-ups and never lets a superseded query replace follow-up memory. Seven
more layers expose analyst records.
- Tool results carry a code-built say line and a display card; the voice card in
the mic control shows captions, plan steps and the result
(src/voice/resultDisplay.js is the one adapter). Slow tools narrate their
current step out of band; holding Space while the assistant talks cuts it off.
Turn spans go to the debug log; caption transcription is metered; the log
omits what was said unless GEV_VOICE_LOG_CONTENT=1.
- Point and ask: the Space-down pointer snapshot resolves "this", "that one" and
"around here" through 'pointer' targets, and "the second one" through
referent:n from the last numbered list.
- Prompt states each policy once (instructions + tools 48.5k -> 46.7k chars).
- Voice benchmark harness (scripts/qa-voice-bench.mjs, scripts/voice-bench/)
shares its phrase table and grading with scripts/qa-voice-routing.mjs.
Supersedes #235 by @cora-fresh-labs, ported onto the current component layout with fixes from review.
Verified before merge: format, package boundaries, unit suite, build, and the browser tracking suite; a browser harness for the change where one applies; adversarial review applied.
Future work (not in this PR): hand-drawn marks do not persist across a reload today, and neither do spoken annotations; there is no storage layer for marks. The natural home is Scenes: a drawing that belongs to a shot and persists per shot, so a scene animation can carry its own marks. Needs an owner.
Co-authored-by: Clint Borle <[email protected]>
Bilawal Sidhu