05-multi-domain-comprehensive.yaml 2.2 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071
  1. id: openagent-contextscout-multi-domain
  2. name: "OpenAgent: Multi-Domain - ContextScout Finds All Relevant Context"
  3. description: |
  4. Tests that OpenAgent uses ContextScout for complex, multi-domain tasks
  5. that require multiple context files from different areas.
  6. This validates:
  7. - Agent recognizes multi-domain complexity
  8. - Agent delegates to ContextScout for comprehensive discovery
  9. - ContextScout finds ALL relevant files (not just one)
  10. - Agent loads multiple context files in priority order
  11. - Agent synthesizes information from multiple sources
  12. Expected Behavior:
  13. - SHOULD delegate to ContextScout
  14. - MUST find multiple relevant context files:
  15. * Agent creation guide
  16. * Code standards
  17. * Test standards
  18. * Eval framework concepts
  19. - SHOULD load files in priority order
  20. - SHOULD provide comprehensive answer
  21. This test SHOULD FAIL if:
  22. - Agent doesn't use ContextScout
  23. - ContextScout only finds partial context (misses key files)
  24. - Agent loads files in wrong order (low priority first)
  25. - Agent provides incomplete answer
  26. category: developer
  27. agent: core/openagent
  28. prompts:
  29. - text: |
  30. I want to create a new agent with proper tests and documentation.
  31. What do I need to know? What files should I look at?
  32. approvalStrategy:
  33. type: auto-approve
  34. behavior:
  35. mustUseTools:
  36. - task # Must delegate to ContextScout
  37. - read # Must read multiple discovered files
  38. minToolCalls: 5 # ContextScout + multiple reads
  39. maxToolCalls: 30
  40. expectedViolations:
  41. - rule: approval-gate
  42. shouldViolate: false
  43. severity: error
  44. timeout: 120000 # Multi-domain discovery takes longer
  45. tags:
  46. - openagent
  47. - contextscout-integration
  48. - multi-domain
  49. - comprehensive
  50. - critical
  51. # Expected outcome:
  52. # - Agent delegates to ContextScout
  53. # - ContextScout finds multiple files:
  54. # * .opencode/context/openagents-repo/guides/adding-agent.md
  55. # * .opencode/context/openagents-repo/core-concepts/agents.md
  56. # * .opencode/context/core/standards/code-quality.md
  57. # * .opencode/context/core/standards/test-coverage.md
  58. # * .opencode/context/openagents-repo/core-concepts/evals.md
  59. # - Agent loads all discovered files
  60. # - Agent provides comprehensive guide with all necessary information