Repository navigation

CI: behavioral scenario tests on skill PRs (12 fixtures, Gemini executor) - #10

Open

harshaneel wants to merge 8 commits into

mainfrom

ci/scenario-tests

Open

CI: behavioral scenario tests on skill PRs (12 fixtures, Gemini executor)#10
harshaneel wants to merge 8 commits into
mainfrom
ci/scenario-tests

Commits on Jul 10, 2026
- CI: run the behavioral scenario fixtures as a PR workflow
  
  harshaneel
  and
  claude
  committed
  9af6625
- Scenario tests: widen scenario-3 verdict, use non-lite flash executor
  
  harshaneel
  and
  claude
  committed
  40937cc
- Scenario tests: parser fix, min-pass gate, fixture correction
  
  harshaneel
  and
  claude
  committed
  b0aaa13
- Scenario tests: back to flash-lite executor, gate at measured baseline
  
  harshaneel
  and
  claude
  committed
  e94e658
- Scenario tests: best-of-2 per scenario, revert temperature experiment
  
  harshaneel
  and
  claude
  committed
  ec4423d
- Scenario tests: two-pass execution (draft, then skill-driven revision)
  
  harshaneel
  and
  claude
  committed
  58bb9d6
- Scenario tests: pin gate at measured executor floor, label it as such
  
  harshaneel
  and
  claude
  committed
  40ffe78
- Scenario tests: route ai-check scenarios to the non-lite flash tier
  
  harshaneel
  and
  claude
  committed
  cd2b260