SIM A-B4-2 / B4

battery · batteryA/B4-2/result.json

← back to index
REVIEW: AMBER FAIL
Assertion quality: strong — this is the variant that found a real product bug, which is exactly what a variant is for
Score: 3.0
variant: invalid-segment-rejectionbuilt: yes
6 turns 0 ⛔ hard-rule violations 1 assertion failures

No ⛔ hard-rule violations reported.

Assertion failures (fixture/spec assertions that failed — they lower the score, they are not red lines):

  • an unknown mentor name is rejected, not printed as a label: exit 0 — NOT VALIDATED: --dry-run prints 'mentor=notarealmentor' as a timeline label and exits 0, so the dry-run timeline the battery certifies can name a voice that does not exist. output: dur(s) cum(s) 0 mentor mentor=notarealmentor 1.0 1.0 TOTAL estimated: 1.0s (0.02 min)

Notes: FAIL

systemfixture runner · B4 — Pre-call briefing audio · run 2 · variant=invalid-segment-rejection · built='yes' · sandbox /var/folders/jf/t4cqx65j5vs9wbm1y5thcv380000gq/T/soh-battery-nx3iu__q/B4-2-z3x3ejfb
coachPASS · build_audio.py exists /Users/agency/Documents/Agty/telegram coach/plans/system-reset-2026-07-25/scripts/build_audio.py
mattFAIL · an unknown mentor name is rejected, not printed as a label exit 0 — NOT VALIDATED: --dry-run prints 'mentor=notarealmentor' as a timeline label and exits 0, so the dry-run timeline the battery certifies can name a voice that does not exist. output: dur(s) cum(s) 0 mentor mentor=notarealmentor 1.0 1.0 TOTAL estimated: 1.0s (0.02 min)
coachPASS · an unknown segment type is rejected exit 1; y/Documents/Agty/telegram coach/plans/system-reset-2026-07-25/scripts/build_audio.py", line 151, in build raise ValueError(f"unknown segment type {typ}") ValueError: unknown segment type wormhole
coachPASS · an empty segment list is handled explicitly # type detail dur(s) cum(s) TOTAL estimated: 0.0s (0.00 min)
system3/4 assertions passed (0 skipped as not-applicable) → raw score 3.0/4