04-multi-domain-comprehensive.yaml 2.5 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687
  1. id: contextscout-multi-domain
  2. name: "ContextScout: Multi-Domain Discovery - Find All Relevant Files"
  3. description: |
  4. Tests that ContextScout can discover multiple relevant files across
  5. different domains when given a complex, multi-faceted query.
  6. This validates:
  7. - ContextScout searches multiple directories
  8. - ContextScout finds files from different categories
  9. - ContextScout prioritizes correctly (critical first)
  10. - ContextScout doesn't miss important files
  11. - ContextScout provides comprehensive results
  12. Expected Behavior:
  13. - MUST find multiple relevant files:
  14. * Agent creation guide
  15. * Code standards
  16. * Test standards
  17. * Eval framework concepts
  18. - MUST prioritize correctly (critical > high > medium)
  19. - SHOULD provide loading strategy
  20. - SHOULD include line ranges for each file
  21. - MUST NOT miss critical files
  22. This test SHOULD FAIL if:
  23. - ContextScout only finds partial results
  24. - ContextScout prioritizes incorrectly
  25. - ContextScout misses critical files
  26. - ContextScout returns files in wrong order
  27. category: developer
  28. agent: subagents/core/contextscout
  29. prompts:
  30. - text: |
  31. Search for context files related to: creating a new agent with tests
  32. Task type: agent-creation
  33. Return:
  34. - Exact file paths with line ranges
  35. - Priority order (critical, high, medium)
  36. - Key findings from each file
  37. - Loading strategy
  38. Focus on:
  39. - Agent creation process
  40. - Code standards
  41. - Test standards
  42. - Eval framework
  43. approvalStrategy:
  44. type: auto-approve
  45. behavior:
  46. forbiddenTools:
  47. - task
  48. - bash
  49. - write
  50. - edit
  51. minToolCalls: 1 # Multiple discoveries and reads
  52. maxToolCalls: 10
  53. expectedViolations:
  54. - rule: approval-gate
  55. shouldViolate: false
  56. severity: error
  57. timeout: 90000
  58. tags:
  59. - contextscout
  60. - multi-domain
  61. - comprehensive
  62. - discovery
  63. - critical
  64. # Expected outcome:
  65. # - Finds multiple files across domains:
  66. # 1. ⭐⭐⭐⭐⭐ .opencode/context/openagents-repo/guides/adding-agent.md
  67. # 2. ⭐⭐⭐⭐⭐ .opencode/context/openagents-repo/core-concepts/agents.md
  68. # 3. ⭐⭐⭐⭐ .opencode/context/core/standards/code-quality.md
  69. # 4. ⭐⭐⭐⭐ .opencode/context/core/standards/test-coverage.md
  70. # 5. ⭐⭐⭐⭐ .opencode/context/openagents-repo/core-concepts/evals.md
  71. # - Returns structured response with all files
  72. # - Provides loading strategy: "Load guides first, then standards, then concepts"
  73. # - Includes line ranges for key sections in each file