WHAT VEDA IS MAKING FOR YOU · SOURCE-BACKED OR IT SAYS “UNAVAILABLE”
CAPACITYunavailable — no source supplied
OBJECTIVE
No canonical build trace connected — current build state unavailable.
PROGRESSunavailable — no source supplied
ETA
unavailable — no source supplied
BUILT · NOT DEPLOYEDbuilt 0 · deployed 0
Two runtimes, one seam: work lands to the integration branch, which is not the live runtime. Crossing that seam is a separate gate — a self-deploy battery plus your approval. Nothing has left the device.
OWNER GUARANTEES
✓Each item is claimed and landed exactly once.
✓A seat that dies has its work auto-reclaimed.
✓Each build runs in an isolated branch and cannot escape it.
Values live only when the trace is connected; otherwise unavailable.
AGING · QUIET IS A SIGNAL
Last completed: unavailable — no source supplied. Exact elapsed unavailable — no completion source is connected.
RECEIPTS
Verdict recordchecks + verify · unavailable
Events-trail linecustody · unavailable
Landed commit shaunavailable
The whole custody trail is the events.jsonl line history — available once the trace is wired.
SIDE STATES{{ x }}live positions appear only when the trace is connected
THE WORK
CURRENTunavailable — no source supplied
unavailable — no source supplied
unavailable — no source supplied
NEXT · BLOCKEDneeds owner decision
Resolve disputed place
Canal vs Sakaimachi on the Otaru day. Held for your call — Veda won’t guess.
MODULE BELTthe beats, moving through
{{ m.mark }}{{ m.name }}
{{ m.label }}
CHECKS · GATES OUTRANK MODEL OPINION
{{ c.name }}
{{ c.detail }}
{{ c.tag }}
SEATS · LIVE ROSTER FROM THE TRACE
{{ s.name }}{{ s.cap }}{{ s.model }}
{{ s.activity }}
roster unavailable — no trace connected
One row per seat the trace reports — whatever the names, however many (elastic 1–12) — each with its capability (author / reviewer) and current activity; a reviewer always reviews another seat’s item. A model id shows only when resolved, else model resolved at install.
THE GATE · GATES OUTRANK MODEL OPINION
▸The objective check (verify_cmd) decides — a red check overrides any model PASS.
▸Only the claim-holding reviewer rules — never the seat that wrote the item (self-review is impossible).
▸A down or unreadable reviewer releases the item for another — no verdict recorded, no attempt spent (fail-closed; never auto-approved).
THREE ATTENTION CLASSES · OWNER · REASON · NEXT ACTION · NEXT CHECK
{{ b.name }}
{{ b.desc }}
ownerreasonnext actionnext check
all unavailable
ALTERNATE — NOT CURRENT · A DISPUTED PLACE READING. THE EVIDENCE IS SHOWN; A GATE RESOLVES IT, NOT THE “SMARTER” MODEL.
THE LOW-CONFIDENCE OTARU PLACE — WHERE WAS THIS TAKEN?
READING APOINTS TO
Otaru Canal
EVIDENCE · ROUTE TRACE
The route trace reads as the Canal.
VS
READING BPOINTS TO
Sakaimachi Street
EVIDENCE · PHOTO #9
Photo #9 shows Sakaimachi signage and stone storefronts.
DETERMINISTIC RULE
The gate decides — not a model.
Reviewers don’t argue: the objective verify_cmd gate and the verdict boundary decide, and a down or unreadable reviewer releases the item for another (no verdict, no attempt spent). This place ambiguity is a low-confidence inference held for your evidence call — the two readings are shown, neither is applied.
YOUR CALL · THE ONLY WAY THIS RESOLVES
Until you decide, the place stays an open gap. Nothing is assumed, blended, or shipped.
Trust — the build floor
Percentages and ETAs appear only when a real source backs them. Otherwise you’ll read “unavailable.”
Built is not deployed. Nothing releases without the exact-artifact gate and your approval.
Models are named only where they’re accountable for real work. When they disagree, a gate decides — never “the smarter model.”