Problem
Expected adversarial oracle findings produce nonzero nested fuzz-suite executions. The recipe-wide verification scan later treats every nonzero execution as an infrastructure failure, so a campaign that successfully discovers, minimizes, seals, and replays declared findings exits as a failed recipe.
Expected
Represent declared adversarial findings distinctly from orchestration/runtime failures. Finding executions must remain visible and nonzero in campaign evidence, while actual campaign incompleteness, runtime errors, cancellation, cleanup failure, and unrelated failed recipe steps continue to fail closed.
Context
Confirmed while completing #2017 after merged PR #2028. The fix should remain generic and preserve existing recipe verification semantics outside adversarial campaign phases.
Refs #2017
Problem
Expected adversarial oracle findings produce nonzero nested fuzz-suite executions. The recipe-wide verification scan later treats every nonzero execution as an infrastructure failure, so a campaign that successfully discovers, minimizes, seals, and replays declared findings exits as a failed recipe.
Expected
Represent declared adversarial findings distinctly from orchestration/runtime failures. Finding executions must remain visible and nonzero in campaign evidence, while actual campaign incompleteness, runtime errors, cancellation, cleanup failure, and unrelated failed recipe steps continue to fail closed.
Context
Confirmed while completing #2017 after merged PR #2028. The fix should remain generic and preserve existing recipe verification semantics outside adversarial campaign phases.
Refs #2017