
== ToolScale deepseek-v4-pro: 112 successful, 172 failed runs, 9.3 s
   successful runs flagged   figured   1.8%  baseline   3.6%
   failed runs flagged       figured  36.0%  baseline  65.7%
   corruptions caught        figured 307/310  baseline 310
   substitutions caught      figured 0/114  baseline 0
   calls to unknown tools    successful runs 0  failed runs 0
   ambiguous_before asks     successful   1.9% of writes, failed  47.5%
   confirm_before asks       successful   7.7% of writes, failed  55.6%

== ToolScale Qwen-3.6-plus: 285 successful, 297 failed runs, 17.0 s
   successful runs flagged   figured   1.1%  baseline  14.4%
   failed runs flagged       figured   1.0%  baseline  19.9%
   corruptions caught        figured 773/774  baseline 772
   substitutions caught      figured 1/310  baseline 0
   calls to unknown tools    successful runs 0  failed runs 0
   ambiguous_before asks     successful   2.7% of writes, failed   0.0%
   confirm_before asks       successful   2.7% of writes, failed   0.0%

== speed: 3899 calls, before_call p50 39.0 µs, p99 341.0 µs
