You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
PR #123 review-leftover: Cases such as native-subagent-delegates-further (line 9), `workflow-without-op [faf8dcc8374e] #125
evals/mcp-tool-use/cases/orchestration.ndjson:9 Cases such as native-subagent-delegates-further (line 9), workflow-without-opt-in (line 6) and sol-native-agent-type-instead-of-delegate (line 7) still test native-subagent or Workflow behavior that Claude and Codex sessions can no longer reach. Since the eval now removes those tools, these cases pass without exercising the behavior they were intended to test.
Activity