Sprint Status — 95% Confidence Push

Auto-regenerated every daemon tick. Open in Obsidian for the live view.

Headline

1240 / 2640 cells done (47.0%) — 1400 remaining

Velocity: (insufficient history; need 1+ hour of dashboard runs)

Cell state distribution

StateCountCounts as
confidence: high433✅ done
populated n=0 (declared gap)807✅ done
confidence: medium327⏳ needs more sources
confidence: low484⏳ needs more sources
pending / missing570🔴 needs first research
other / unknown19🟡 audit

Pipeline health

ComponentState
Daemon heartbeat⚠️ 138 min ago (stale)
auto-fetched/ PDFs in queue221 (fresh work for Codex/daemon ingest)
Audit queue: pending5 ✅
Audit queue: pending-audit-spawn-failed33 (async lane)
Wishlist: total downloaded historically99
Wishlist: total ingested historically0
Repo sync⚠️ 17 commit(s) ahead of origin/main
Commits last 1 hour14
Commits last 24 hours122

Most-needy ingredients (top 10 by total source-delta)

SlugSource-delta needed
baby-food100
beverages100
black-pepper100
bread100
breast milk100
carrots100
casein100
cashew-nuts100
cashews100
cereal-bars100

Last commit

8a9715c 5 seconds ago — audit-queue: tinggi2025-spices-herbs-queensland audited-revised

How to read this

The headline number is the only one that matters: cells done / total. Everything else is process metric.

A well-oiled engine looks like:

  • Heartbeat fresh (<35 min old, daemon ticking)
  • auto-fetched/ depth fluctuating (refill then drain then refill — not stuck at zero or piling up indefinitely)
  • Audit queue pending ≤5 (Codex draining it)
  • Commits/hour ≥ 5 during active sprint
  • Velocity positive (cells done going UP between dashboards)
  • Unpushed = 0 (daemon pushes every tick)

When velocity stalls for >2 hours, something upstream broke. Check data/evidence/autonomy/daemon-runs/ for the most recent tick log and look for “FAIL:” lines.