Statistics

Aggregated from the poller's snapshot (updated 4m ago). Comment fates exist since posting went live; verify verdicts & activity only for traced reviews (since 2026-07-05) — missing history shows as honest «no data».

findings posted

0 got textual replies

resolved rate

—

0 resolved · 0 in discussion · 0 open

$ / resolved finding

—

total spend $0.00

verify refuted rate

no data

0 refuted · 0 real · 0 uncertain

coverage & funnel

0 reviews

0 → 0 raw→kept

0 skipped · 0 errors

Finding fate by severity

severity	posted	resolved	resolved share

Resolved-rate by confidence tier is the calibration check — a high-severity row resolving less often than minor is a red flag.

Finding fate by category

category	posted	resolved	resolved share

Confidence calibration

Resolved-rate by the model's own confidence. BY DESIGN only high-confidence findings are posted as inline comments (medium/low go into the summary, with no individual fate) — so this table can only ever contain the high tier; it still shows whether «high» earns its name.

confidence	posted	resolved	resolved share

Implicit fix-rate (re-reviewed PRs)

no data — needs PRs re-reviewed after new pushes

Skeptic vs humans

no data — verify verdicts exist only for traced reviews (since 2026-07-05)

AST digest ROI (brain activity)

no data — brain activity is recorded since 2026-07-05

Ticket alignment (reviews)

Per project

project	reviews	kept findings	posted	resolved	cost