description precision report
tests.json + results-2026-03-14.md · baseline vs skill output · per-case eval viewer
mode
static
dynamic
test case
report view
selected case
aggregate
loading eval artifacts...
live generation
generate live
replay current run
dynamic mode sends your prompt to a live model run, then draws baseline, skill output, and evals as they come back.
generation timeline
baseline
waiting
skill output
waiting
eval draw
waiting
static artifact mode
precision / voice / coherence space
eval summary
reading notes
A — baseline
B — skill
showing sentence-by-sentence output from tests.json
metrics and narrative are parsed from results-2026-03-14.md