Repository navigation

feat(experiments): controlled baseline-vs-candidate eval harness for … - #307

Open

Matthew (theworker02) wants to merge 3 commits into

microsoft:mainfrom

theworker02:devin/skillopt-eval-harness

Open

feat(experiments): controlled baseline-vs-candidate eval harness for …#307
Matthew (theworker02) wants to merge 3 commits into
microsoft:mainfrom
theworker02:devin/skillopt-eval-harness

Commits on Oct 7, 2026
- feat(experiments): controlled baseline-vs-candidate eval harness for #132
  
  Matthew (theworker02)
  and
  devin-ai-integration[bot]
  committed
  ce4e6f7
Commits on Oct 13, 2026
- fix(eval): preserve invocation evidence and fail closed
  Matthew (theworker02)
  committed
  b50a9f7
- style: remove trailing blank line
  Matthew (theworker02)
  committed
  c653e0e