02-delegation-test.yaml 1.8 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465
  1. id: golden-02-delegation-test
  2. name: "Golden 02: Delegation Test - Parent & Child Agent"
  3. description: |
  4. Tests multi-agent delegation workflow.
  5. This test validates:
  6. - Agent can delegate to a subagent using task tool
  7. - Subagent receives context and responds correctly
  8. - Multi-agent logging captures both sessions
  9. - Parent-child communication works
  10. Expected completion time: 25-35 seconds (includes LLM inference for both agents)
  11. This is more complex than the minimal smoke test because:
  12. - Tests delegation (core feature of agent system)
  13. - Runs TWO agents (parent + child)
  14. - Tests context passing
  15. - Validates multi-agent logging system
  16. category: developer
  17. prompts:
  18. - text: |
  19. Call the simple-responder subagent and ask it to respond. Use the task tool to delegate.
  20. - text: |
  21. Yes, proceed with the delegation.
  22. delayMs: 2000
  23. approvalStrategy:
  24. type: auto-approve
  25. behavior:
  26. mustUseTools: [task]
  27. minToolCalls: 1
  28. maxToolCalls: 10
  29. requiresApproval: false # This is a simple test, no approval ceremony needed
  30. expectedAgent: "openagent" # Validate correct agent is loaded
  31. expectedModel: "opencode/grok-code" # Validate correct model is used
  32. expectedResponse:
  33. contains:
  34. - "AWESOME TESTING"
  35. description: "Subagent should respond with 'AWESOME TESTING'"
  36. expectedViolations:
  37. - rule: approval-gate
  38. shouldViolate: false
  39. severity: error
  40. description: "Should request approval before delegation"
  41. - rule: context-loading
  42. shouldViolate: false
  43. severity: warning
  44. description: "Should load delegation context"
  45. - rule: delegation
  46. shouldViolate: false
  47. severity: error
  48. description: "Should successfully delegate to subagent using task tool"
  49. timeout: 90000
  50. tags:
  51. - golden
  52. - delegation
  53. - subagent
  54. - multi-agent
  55. - parent-child